an... .
1.“? n 2.....
. . 3 y . 2F J-ucF.‘

!
.6? .
.. .. .
um . In?!

 

.21... a»

:12 .3......:

.3 2.1.1.5.“. .
. .52.»... .i

Whig) '

ii

2
. &
....§.......n a. !.

 

amgmwﬁéggg

I'

 

2! ‘ LIBRARY
4‘” Michig'cm State
University

 

This is to certify that the
dissertation entitled

g AGROECOSYSTEM DIVERSITY: IMPACTS ON WEED
“I COMMUNITIES AND CROP YIELD

presented by

 

Richard Garner Smith

has been accepted towards fulﬁllment
of the requirements for the

Doctoral degree in Plant Biology and Program in

Ecology. Evolutionary Biology &
Behavior

27544” :1. IL)

Major Professor’s Signature
5’] 2.3 / 05"
/ /

Date

 

MSU is an Afﬁrmative Action/Equal Opportunity Institution

 

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

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 c-JcTﬁcmateoqum-ms

AGROECOSYSTEM DIVERSITY: IMPACTS ON WEED COMMUNITIES AND
CROP YIELD

By

Richard Garner Smith

A DISSERTATION

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

DOCTOR OF PHILOSOPHY
Department of Plant Biology
Program in Ecology, Evolutionary Biology & Behavior
W. K. Kellogg Biological Station

2005

ABSTRACT

AGROECOSYSTEM DIVERSITY: IMPACTS ON WEED COMMUNITIES AND
CROP YIELD

By
Richard Garner Smith

Ecological theory regarding the role of species diversity in the functioning of
terrestrial ecosystems may have important implications for the design and management
of agricultural systems that are less reliant on inputs of synthetic chemical fertilizers
and pesticides for fertility and pest control. However, much of the evidence supporting
the importance of diversity in ecosystem function comes from grassland systems where
biotic conditions and disturbance regimes differ from those in most managed systems.
Furthermore, agricultural studies that purport to demonstrate the impacts of crop
diversity on crop yields and pest regulation are often confounded with applications of
chemical fertilizers and pesticides. My research aims to understand how ecological
theory regarding the role of diversity in ecosystem function can be used to predict weed
and crop yield response to increasing crop diversity.

I examined the relationships between row-crop diversity and ecosystem
functions related to weed population regulation and crop productivity in two long—tenn
experimental agroecosystems. The ﬁrst system compared corn grown in continuous
monoculture and in rotation with soybean and winter wheat under two input
management systems: conventional and organic-based. The second system was an
experimental manipulation of cr0p diversity with no external chemical inputs of

fertilizer or herbicides. Crop diversity was manipulated by growing corn, soybean, and

winter wheat continuously and in two and three-crop rotations and with either no, one,
or two cover crops annually.

Weed communities and crop yield both varied in response to these treatments. In
the ﬁrst system, weed communities were most variable from year to year in the organic
rotation, while corn yield variability was highest in the least diverse cropping system
(conventional monoculture). In the diversity experiment, crop diversity had relatively
little effect on the abundance or composition of weeds. Cover crops in wheat had
strong suppressive effects on weeds in both continuous wheat grown in monocultures
and in rotation with corn and soybean, while cover crops had little effect on weed
abundance or species composition in corn and soybean.

In contrast to the effect on weed communities, crop diversity had large effects
on crop productivity. Corn yields varied across the diversity treatments and in the most
diverse treatments were comparable to those in conventionally managed systems in the
region. The positive effect of crop diversity on yields in corn was driven by increased
soil N that was related to the number of legume species in the rotation (red clover and
soybean). Soybean yields were also higher in rotations with higher crop diversity,
while winter wheat yields were unaffected by the number of crops in the rotation.

These results suggest that increasing crop diversity can have signiﬁcant impacts
on crop yields, particularly in corn, and may decrease the need for intensive chemical
management by suppressing weeds during phases of the rotation that are conducive to
high cover crop growth (i.e. winter wheat) and by increasing soil nitrogen and the

competitive advantage of crops over weeds.

Dedicated to my late grandparents, George L. Smith and Irene E. Smith

iv

ACKNOWLEDGMENTS

This work could not have been accomplished without the help of many people.
Foremost, I would like to thank my major advisor, Dr. Katherine Gross, for her
patience, thoughtful advice, ecological insight, and friendship throughout my graduate
career. My professional and personal life has been greatly enriched by our interactions.
I would also like to thank the past and present members of my graduate committee: Dr.
Richard Harwood, Dr. Doug Landis, Dr. Carolyn Malmstréim, Dr. Gary Mittelbach, and
Dr. Doug Buhler, for guidance over the past six and a half years.

My ﬁeld work was made much easier, and inﬁnitely more enjoyable, by the
assistance and dedication of Laura Bell, Joey Dembs, Ryan Disney, Matt Gorentz,
Cinnamon Hayes, Jonathan Hernandez, Stephanie Januchowski, Beth Jasiak, Chelsea
Leroy, Mark Mittelbach, John Mittelbach, Jenny Monison, Pam Moseley, Nate
Rathbun, Justin Rensch, Kendra Sales, Chris Start, Lee Sutkowi, C.J. Wyckoff, Pete
Wyckoff, and Margaret Yancey. I would especially like to acknowledge the assistance
of Brandon Young, whose life enriched the lives of all who knew him, and who will be
greatly missed.

My research beneﬁted from the help, advice, and friendship of Carol Baker.
Mark Hammond and Stacey Andres also contributed to the project. Joe Simmons and
Greg Parker were a pleasure to work with and expertly maintained the experimental
treatments in the Biodiversity Experiment Plots (BEP) and the Living Field Laboratory

(LFL), respectively.

The KBS has been a great place to live and work, largely due to the outstanding
efforts of Nina Consolatti, John Gorentz, Alice Gillespie, Sally Shaw, Michelle
Bosanoz, Janell Lovan, and Melissa Yost. The resident faculty, Dr. Tom Getty, Dr.
Steve Hamilton, Dr. Jeff Connor, and Dr. Phil Robertson, make the KBS a stimulating
academic (and otherwise) environment.

The greatest part about being a graduate student at the KBS has been the friends
I have made. Sarah Emery and Erica Garcia made lunchtime, trips to campus, and the
daily grind positively enjoyable. Greg Houseman, Heather Sahli, Terry Loecke, and
Jeremy Wojdak could always be counted on for conversation, commiseration, and
camaraderie. The present and past KBS’ers, Stuart Grandy, Meg Duffy, Angela Roles,
Amanda Kurzman, Frances Knapczyk, Amy Burgin, Nicole Reid, Wendy Mahaney,
Katie Lander, Emily Grrnan, Sara Parr, James DeYoung, Drew Corbin, Tara Darcey-
Hall, Spencer Hall, Claire McSwiney, Claes Becker, Lena Sivars-Becker, Chad Brassil,
Eric Thobaben, Nate Dom, Skip VanBloem, Stephanie Whitrnire, and Laura Broughton
have all been wonderful ﬁiends and colleagues.

Above all I would like to thank my wife Angela Smith who is and always will
be the inspiration for everything I do. Without her unfaltering support and love, none of

this would have been possible.

vi

TABLE OF CONTENTS

LIST OF TABLES ................................................................................. ix
LIST OF FIGURES ................................................................................ xi
CHAPTER 1
INTRODUCTION .................................................................................. 1
Study sites ............................................................................................. 3
Living ﬁeld laboratory .............................................................................. 3
Biodiversity experiment plots ...................................................................... 6
Questions of interest .............................................................................. 17
Organization of dissertation ...................................................................... 19
Literature cited ....................................................................................... 20
CHAPTER 2
EVIDENCE OF RAPID SEED BANK DYNAMICS IN ROW CROPS .................. 24
Abstract ............................................................................................. 24
Introduction...................................................... .................................. 24
Materials and methods ............................................................................. 27
Study sites ................................................................................... 27
Seed bank sampling ...................................................................... 29
Statistical analyses ......................................................................... 29
Results ............................................................................................... 31
LFL Seed bank density and species richness .......................................... 31
LF L Community abundance and composition ........................................ 32
BEP Seed bank dynamics ................................................................ 36
Discussion ........................................................................................... 39
Acknowledgments ................................................................................. 41
Literature cited ..................................................................................... 43
CHAPTER 3
WEED COMMUNITY AND CROP YIELD VARIABILITY IN DIVERSE CORN
AGROECOSYSTEMS ........................................................................... 46
Abstract ............................................................................................. 46
Introduction ........................................................................................ 47
Materials and methods ............................................................................. 49
Study site .................................................................................. 49
Sampling. .................................................................................. 50
Calculation of variability ................................................................. 51
Statistical analyses ........................................................................ 52
Results and discussion ............................................................................ 52
Weed community biomass and richness ............................................... 53
Weed community variability ............................................................ 59

vii

Corn yields and variability .............................................................. 62

Potential explanations and implications ............................................... 64
Acknowledgments ................................................................................. 66
Literature cited ..................................................................................... 67
CHAPTER 4
WEED COMMUNITY RESPONSE TO INCREASING CROP DIVERSITY. . . . .......69
Abstract .............................................................................................. 69
Introduction ........................................................................................ 70
Materials and methods ............................................................................ 73

Study site ................................................................................... 73

Diversity treatments ....................................................................... 73

Vegetation sampling ...................................................................... 76

Soil resources and light .................................................................. 76

Statistical analyses ........................................................................ 77
Results .............................................................................................. 80

Weed biomass ............................................................................. 80

Species richness and diversity .......................................................... 83

Community composition and structure ................................................ 86

Soil and light resources and relationships to the weed community ................ 93
Discussion ......................................................................................... 101

. Acknowledgments ................................................................................ 107
Literature cited .................................................................................... 109
CHAPTER 5
CROPPING-SYSTEM DIVERSITY AFFECTS ECOSYSTEM FUNCTION IN
ROW CROPS ..................................................................................... 114
Abstract ............................................................................................ 114
Introduction ....................................................................................... 1 14
Materials and methods ........................................................................... 116

Study site ................................................................................. 116

Sampling ................................................................................. 118

Statistical analyses ...................................................................... 119
Results and discussion ........................................................................... 119
Acknowledgments ............................................................................... l 26
Literature cited .................................................................................... 127
CHAPTER 6
CONCLUSIONS AND FUTURE DIRECTIONS ........................................... 130
Future directions ................................................................................. 131
Literature cited .................................................................................... 134
APPENDICES .................................................................................... l 36
APPENDIX A: CHAPTER 4 APPENDIX ................................................... 137

viii

LIST OF TABLES

Table 1.1. Crop diversity treatments at the KBS LTER Biodiversity Experiment Plots
(BEP). Treatments were established in 2000 and include monocultures and two- and
three-crop rotations of corn (C), soybean (S) and wheat (W), with and without grass or
legume cover crops. NA, not applicable............................................... ......... 8

Table 2.1: Relative density (%) of weed species that emerged from the soil seed bank
samples collected in May 2001 in continuous maize (CM, never planted to winter
wheat) and the rotational phase of maize (RM, last planted to wheat in 1998) and winter
wheat (RW, last planted to wheat in 2000) phases of the rotation ........................... 34

Table 2.2: Relative densities (%) in the soil seed bank of the eight weed species
common to both the continuous winter wheat monoculture and the wheat rotation in
2001 and following planting of maize (2002) and soybean (2003) ......................... 37

Table 3.1: Weed biomass, species richness, and crop yields (mean i SE; n = 4) in corn
(2001-2004) in conventional (CONV) and organic (ORG) input systems in the KBS
LF L. Rotations are continuous (CC) and rotated (R) com (see text) ....................... 55

Table 3.2: Weed species abundance in conventional (CONV) and organic (ORG) input
systems under continuous (CC) and rotation (R) production. Values are mean % of total
weed biomass and CV averaged across years (2001-2004) .................................. 58

Table 4.1: Crop diversity treatments at the KBS LTER Biodiversity Experiment Plots
(BEP). Treatments were established in 2000 and include monocultures and two- and
three-crop rotations of corn, soybeans and wheat, with and without grass and legume
cover crops .......................................................................................... 75

Table 4.2: Results of indicator species analysis (ISA) on weed community composition
(presence/absence) at the Biodiversity Experiment Plots. ISA Indicator Value (IV)
calculated from annual biomass harvest averaged over three years (2002-2004). Bold
type indicates signiﬁcant associations with a crop species at the P < 0.05 level of
probability .......................................................................................... 90

Table 4.3: Results of distance-based multivariate analysis of variance (Anderson 2001,
McArdle and Anderson 2001) of the effects of blocking (B), crop diversity (D), and
year (Y) on weed abundance and composition from 2002 through 2004 in the KBS
Biodiversity Experiment Plots. ................................................................... 91

Table 4.4: Pearson correlation coefﬁcients and R2 between environmental and

biological variables and axes scores from NMDS ordination analysis of weed
abundance and composition 2004 in the KBS Biodiversity Experiment Plots .......... 100

ix

Table 4.5: Results of distance-based multivariate multiple regression with forward
selection (McArdle and Anderson 2001) on weed abundance and composition in 2004
in the KBS Biodiversity Experiment Plots .................................................... 102

Table 5.1: Repeated measures ANOVA for the effects of cropping system diversity on
crop yields and weed biomass in the KBS Biodiversity Experiment Plots (BEP) over
three years (2002-2004). Signiﬁcant effects are bolded for clarity ....................... 121

APPENDIX A

Appendix A4.1: Indicator values (IV) of species found in spring (S) and fall-tilled (F)
fallow treatments from 2002-2004. Nativity indicates whether species were introduced
(1) or native (N) to the USA. Life cycle indicates whether species are annual (A),
biennial (B), or perennial (P). Species in bold are signiﬁcant at P < 0.05 % ............ 137

LIST OF FIGURES

Figure 1.]: Aerial photo of the KBS LTER main site and surrounding landscape.
Arrows indicate study sites. Abbreviations are LFL, Living Field Laboratory; BEP,
Biodiversity Experiment Plots; LTER, Long Term Ecological Research site; KBS, W.
K. Kellogg Biological Station academic campus ................................................ 4

Figure 1.2: Plot map showing locations of 20 of 2] treatments in four replicate blocks
at the Biodiversity Experiment Plots at the Kellogg Biological Station Long Term
Ecological Research site. See Table 1.1 for treatment codes. Treatment 321 is located
on the western edge of blocks 1-3 and on the western edge of plot 420 in block 4........ .9

Figure 1.3: Schematic representation of Treatment Group] (continuous monocultures
of each crop species) in the BEP over the study period. Colored bars indicate length of
time, from crop sowing to harvest, that each species is present in each of the three
treatment sequences. Months are 1, January; 4, April; 7, July; 10, October.. . .. 10

Figure 1.4: Schematic representation of Treatment Group 2 (continuous monocultures
with a cover crop) in the BEP over the study period. Colored bars indicate length of
time, from crop sowing to harvest, that each species is present in each of the three
treatment sequences. Months are 1, January; 4, April; 7, July; 10, October ............... 11

Figure 1.5: Schematic representation of Treatment Group 3 (two-crop rotation) in the
BEP over the study period. Colored bars indicate length of time, from crop sowing to
harvest, that each species is present in each of the three treatment sequences. Months
are 1, January; 4, April; 7, July; 10, October ................................................... 12

Figure 1.6: Schematic representation of Treatment Group 4 (three-crop rotation) in the
BEP over the study period. Colored bars indicate length of time, from crop sowing to
harvest, that each species is present in each of the three treatment sequences. Months
are 1, January; 4, April; 7, July; 10, October ................................................... 14

Figure 1.7: Schematic representation of Treatment Group 5 (three-crop rotation with
one cover crop annually) in the BEP over the study period. Colored bars indicate length
of time, from crop sowing to harvest, that each species is present in each of the three
treatment sequences. Months are 1, January; 4, April; 7, July; 10, October ............... 15

Figure 1.8: Schematic representation of Treatment Group 6 (three-crop rotation with
two cover crops annually) in the BEP over the study period. Colored bars indicate
length of time, from crop sowing to harvest, that each species is present in each of the
three treatment sequences. Months are 1, January; 4, April; 7, July; 10, October.... .....16

Figure 2.1: Density (A) and species richness (B) in May 2001 of seedlings emerged
from the soil seed bank in continuous maize (CM, never planted to winter wheat) and

xi

the rotation phase of maize (RM, planted to winter wheat in 1998) and winter wheat
(RW, planted to winter wheat in 2000) of a four year maize-maize-soybean-winter
wheat rotation under conventional (black bars) and organic-based (white bars) nutrient
and weed management. Bars are means i 1 SE. There was no effect of input system
(organic vs. conventional) so bars sharing the same lower case letter indicate no
signiﬁcant difference among rotation phases at P = 0.05 (Tukey-Kramer Test) ......... 33

Figure 2.2: NMDS ordination of weed seed banks from soils collected in May 2001 in
continuous maize (never planted to winter wheat) (triangles) and the rotation phase of
maize (last planted to winter wheat in 1998) (inverted triangles) and winter wheat (last
planted to winter wheat in 2000) (squares) of a four year maize-maize-soybean-winter
wheat rotation under conventional (ﬁlled symbols) and organic-based (open symbols)
nutrient and weed management. Values plotted are replicate scores. Only the two most
explanatory axes (r2 = 0.779) are displayed .................................................... 35

Figure 2.3: NMDS ordination of weed seed banks in four replicate plots (ﬁlled
symbols) in 2001, 2002, and 2003 following planting to winter wheat (inverted
triangles), maize (squares), and soybean (triangle), respectively, and the corresponding
continuous monoculture in that year. Replicate plots are labeled 1-4 and dashed arrows
indicate each plot’s trajectory through time. Ellipses around monocultures (open
symbols) are for illustrative purposes and do not imply conﬁdence intervals. Only the
two most explanatory axes (r2 = 0.780) are displayed ........................................ 38

Figure 3.1: Mean daily precipitation and air temperature at the KBS LTER for each
month from April-October over the four-year study .......................................... 54

Figure 3.2: Effects of conventional (CONV) and organic (ORG) inputs and crop
rotation on a) weed community biomass from 2001-2004 and b) inter-annual variability
as measured by the coefﬁcient of variation, CV. Error bars are i 1 SE; total biomass, n
= 16; CV of biomass, n = 4. Bars sharing the same letter are not signiﬁcantly different
at the P = 0.05 level (Tukey-Kramer Test) ...................................................... 56

Figure 3.3: Effects of conventional (CONV) and organic (ORG) inputs and crop
rotation on a) weed community species richness from 2001-2004 and b) inter-annual
variability as measured by the coefﬁcient of variation. Error bars are i 1 SE; richness,
n = 16; CV of richness, n = 4. Bars sharing the same letter are not signiﬁcantly
different at the P = 0.05 level (Tukey-Kramer Test) .......................................... 57

Figure 3.4: Effects of conventional (CONV) and organic (ORG) inputs and crop
rotation on a) weed community dissimilarity and b) population-level inter-annual
variability as measured by the coefﬁcient of variation. Error bars are i- 1 SE; n = 4.
Bars sharing the same letter are not signiﬁcantly different at the P = 0.05 level (Tukey-
Kramer Test) ........................................................................................ 61

Figure 3.5: Effects of conventional (CONV) and organic (ORG) inputs and crop
rotation on com a) yields from 2001-2004 and b) inter-annual variability as measured

xii

by the coefﬁcient of variation. Error bars are i 1 SE; yield, n = 16; CV, r1 = 4. Bars
sharing the same letter are not signiﬁcantly different at the P = 0.05 level (Tukey-
Kramer Test) ........................................................................................ 63

Figure 4.1: Weed biomass in relation to crop diversity over time in the KBS
Biodiversity Experiment Plots. Symbols refer to crop diversity treatments (see Table
4.1). Symbols that are shaded indicate treatments with cover crops. Values are means
(3: SE; n = 12) averaged across the three crops ................................................ 81

Figure 4.2: Weed biomass in winter wheat planted under different crop diversity
treatments (2002-2004). Symbols refer to crop diversity treatments (see Table 4.1).
Symbols that are shaded indicate treatments with cover crops. Values are means (:1: SE;
n = 4) ................................................................................................. 82

Figure 4.3: Number of weed species from 2002 through 2004 in crop diversity
treatments at the KBS Biodiversity Experiment Plots. Symbols refer to crop diversity
treatments (see Table 4.1). Symbols that are shaded indicate treatments with cover
crops. Values are means (:1: SE; n =12) averaged across the three main crops .......... 84

Figure 4.4: Number of weed species from 2002 through 2004 in cropping-system
diversity treatments planted to winter wheat. Symbols refer to crop diversity treatments
(see Table 4.1). Symbols that are shaded indicate treatments with cover crops. Values
are means (i: SE; n = 4) ........................................................................... 85

Figure 4.5: Shannon diversity (A) and Simpson’s index of species diversity (B) of the
weed community in winter wheat in 2004 in six cropping-system diversity treatments at
the Biodiversity Experiment Plots. Means (i- SE; n = 4) followed by the same letter are
not signiﬁcantly different at the P < 0.05 level of probability (Tukey HSD test). See
Table 4.1 for treatment codes .................................................................... 87

Figure 4.6: Plot scores for the ﬁrst two most explanatory axes of non-metric
multidimensional scaling analysis of weed composition (presence/absence) from 2002
through 2004 in six cropping-system diversity treatments (Table 4.1) at the Biodiversity
Experiment Plots at the Long Term Ecological Research site at W. K. Kellogg
Biological Station in Hickory Comets, MI. NMDS: stress = 17.992, P < 0.05,
cumulative r2 for axes 2 and 3 = 0.638 .......................................................... 88

Figure 4.7: Plot scores for the ﬁrst two most explanatory axes of non-metric
multidimensional scaling analysis of weed abundance and composition in six diversity
treatments (Table 1) in com (A), soybean (B) and winter wheat (C) in 2004 at the Long
Term Ecological Research site at W. K. Kellogg Biological Station in Hickory Comers,
MI. NMDS, corn: stress = 18.74, P < 0.05, r2 = 0.704; soybean: stress = 16.17, P <
0.05, r2 = 0.583; wheat: stress = 18.95, P < 0.05, r2 = 0.711 ................................. 92

Figure 4.8: Soil inorganic N pools (NO3-N + NH4-N) in treatments planted to (A) com
and (B) winter wheat in six cropping-system diversity treatments (see Table 1) at the

xiii

KBS Biodiversity Experiment Plots. Values are means (i SE; n = 4); bars with the
same letter are not signiﬁcantly different at the P < 0.05 level of probability (Tukey
HSD test). Note different x-axis scales in panels A and B .................................. 95

Figure 4.9: Net N-mineralization (ugN/g dry soil/day) in wheat in six cropping-system
diversity treatments (Table 1) at the Biodiversity Experiment Plots. N-mineralization
rate was measured following 28-day incubation (250C) of soils collected on 16
November 2004. Values are means (i SE; n = 4); bars followed by the same letter are
not signiﬁcantly different at the P < 0.05 level of probability (Tukey HSD test). . . . .....96

Figure 4.10: Soil moisture levels in relation to crop diversity treatments (see Table 4.1)
at the KBS Biodiversity Experiment Plots. Soils were sampled at the end of the
growing season 4 yrs after the treatments were initiated (August 2004). Bars are means
(:1; SE; n = 12) are averaged across the three crops; bars with the same letter are not
signiﬁcantly different at the P < 0.05 level of probability (Tukey HSD test) ............. 98

Figure 4.11: Light levels in winter wheat grown in different crop diversity treatments
on 19 September 2004. Means (i SE; n = 4) followed by the same letter are not
signiﬁcantly different at the P < 0.05 level of probability (Tukey HSD test) ............. 99

Figure 5.1: Crop yields averaged across three years (2002-2004) in the BEP
experiment. Values are means (i .1 SE, n = 12). Dotted lines indicate yields (six year
average) for each crop in a nearby conventionally managed com-soybean-wheat
rotation. The diversity system treatments are: (1) continuous monoculture, (2)
monoculture with one cover crop annually, (3) two crop rotation, (4) three-crop
rotation, (5) three-crop rotation with one cover crop annually, and (6) three-crop
rotation with two cover crops annually ....................................................... 122

Figure 5.2: Relationship between com grain yields (October) and early season

inorganic soil nitrogen (June) in 2004. Refer to ﬁgure 5.1 for a description of the
diversity treatments .............................................................................. 124

xiv

CHAPTER 1
INTRODUCTION

There is growing evidence that diversity plays an important role in the
functioning of terrestrial ecosystems (Chapin et al. 2000, Diaz and Cabido 2001,
Hooper et a1. 2005, Spehn et al. 2005). Community and ecosystem-level processes such
as productivity (Hector et al. 1999, Tilman et al. 2001, Lambers et al. 2004), community
stability (Cottingham et al. 2001) invasibility (Levine and D’Antonio 1999, Naeem et
al. 2000, Fargione and Tilman 2005) and nutrient cycling (Hooper and Vitousek 1998)
have all been shown to vary with species or functional diversity. While the exact
mechanisms by which diversity impacts these functions are often debated (Naeem et al
2000, Huston 1997, Wardle 1999, Loreau and Hector 2001) there is little argument that
the conservation and maintenance of diversity has important implications for ecosystem
services required for human survival (Hooper et al. 2005).

The relationship between diversity and ecosystem function may also have
important implications for the functioning of agricultural systems. The reduction in
cropping-system diversity associated with the intensiﬁcation of agriculture and the
concomitant increase in the need for chemical inputs for fertility and pest control
(Tilman et al. 2002, Robertson and Swinton 2005) suggest that ecosystem processes
related to production agriculture may also be regulated by diversity (Altieri 1999,
Minns et al. 2001). Support for the importance of diversity in agricultural systems
come from a number of recent studies showing that more diverse cropping-systems
often out-perforrn less diverse systems in functions such as weed population regulation

(Liebman and Dyck 1993, Teasdale et al. 2004, Westerrnan et al. 2005), yield quantity

and quality (Vandenneer 1989, Reganold et al. 2001), nutrient cycling (Drinkwater et
al. 1998), and disease resistance (Zhu et al. 2000). Increasing cropping-system diversity
has been advocated as an important tool for the management of weeds and for reducing
reliance on chemical inputs (Dekker 1997, Liebman and Gallandt 1997, Kegode et al.
1999, Liebman and Staver 2001, Westerman et al. 2005).

However, much of the theory and experimental evidence supporting the direct
link between diversity and ecosystem ﬁrnction has been developed in experimental
grassland systems (Hooper et al. 2005), where biotic conditions and disturbance
regimes are very different from those in most agricultural systems (Schlapfer and
Schmid 1999). In addition, a large component of diversity in most agricultural systems
is temporal (crop rotation) (Liebman and Dyck 1993, Liebman and Staver 2001),
meaning that many of the interactions among crop species are indirect, in contrast to
grassland systems where many species interact directly (Tilman 1999). Furthermore,
many agricultural studies supporting the importance of diversity are performed in the
presence of external chemical inputs such as fertilizer and herbicide that can vary with
each crop and may confound the effects of crop diversity on weed abundance and yields
(Liebman and Dyck 1993, Doucet et al. 1999). It is therefore unclear whether and to
what extent the observed relationships between community diversity and ecosystem
function reported in grassland studies can be applied to weed communities and crops in
agroecosystems. The goal of this dissertation was to investigate these relationships in

annually tilled row-crop agricultural systems.

Study sites

The general question I was interested in was: Can ecological theory regarding
the role of diversity in ecosystem function be used to predict weed and crop yield
response to increasing row-crop diversity? Components of this general question were
examined in two long-tenn cropping-system experiments, the Living Field Laboratory
(LFL) and the Biodiversity Experiment Plots (BEP), located at the W. K. Kellogg
Biological Station (KBS) Long Term Ecological Research (LTER) project in
agroecology at Michigan State University in Hickory Comers, MI (Figure 1.1). The
KBS LTER also includes a larger (42 ha) cropping-system manipulation (main site, 7
treatments); however, the purpose of the treatments in the main site study is not to
manipulate crop diversity, but rather chemical inputs and tillage. Therefore the main
site treatments were not sufﬁcient to examine questions related to effects of crop
diversity.

Living Field Laboratory

The studies reported in Chapters 2 and 3 were done at the Living Field
Laboratory (LFL; Sanchez et a1. 2004). The LFL is located within a matrix of
agricultural ﬁelds just northwest from the LTER main site (Figure 1.1). Soils at the site
consist primarily of Kalamazoo silt loam (Typic Hapludalfs) and are 43% sand, 40 %
silt, 17 % clay, have pH of 6.68, and contain 1.08% soil organic carbon (Robertson et
al. 1997). The LP L was established in 1993 by Dr. Richard R. Harwood, then the Mott
Chair in Sustainable Agriculture at MSU, and was designed to evaluate the effects of

extended rotations and different input management systems on crop yield, weed and

 

Figure 1.1: Aerial photo of the KBS LTER main site and surrounding landscape.
Arrows indicate study sites. Abbreviations are LFL, Living Field Laboratory; BEP,
Biodiversity Experiment Plots; LTER, Long Term Ecological Research site; KBS, W.
K. Kellogg Biological Station academic campus.

insect pressure, and soil health. The plots are maintained by the Farm Systems Center
(F SC), which also manages the ﬁelds adjacent to these plots. A heavily grazed horse
pasture borders the LF L on the west and likely is a seed source of perennial wind-
dispersed weeds to the LFL and other annual row crops in this area.

The LFL experiment consists of four input management systems (conventional,
organic-based, integrated fertilizer, and integrated compost) crossed with crop rotation
treatments (continuous corn and each phase of a 4 yr com-com-soybean-winter wheat
rotation). The design is a split-plot in a randomized complete block with four
replications. Input management system is the whole plot factor and rotation is the
subplot factor. In addition, the rotation subplots under organic-based management are
split by a second factor; the western half of each plot is planted with cover crops each
year. However, because the cover crop treatment is not applied randomly within the
rotation subplots, and because there is no cover crop treatment in the conventional input
system, observations in the organic subplots are restricted to either the cover-cropped or
non cover cropped portions of each subplot. Therefore the effective plot size of the
rotation subplots in the organic-based input management systems (with cover crops or
without cover crops) is half that of the conventional input management subplots. All
plots are chisel plowed and ﬁeld cultivated prior to planting. Details regarding the
conventional and organic-based input systems are described below.

The conventional input system receives full inputs of fertilizer, herbicides and
pesticides at rates recommend for the region (Sanchez et al. 2001). Starter fertilizer is
applied to corn at planting and after a pre side-dress nitrate test in June. No fertilizer is

applied to soybeans. Fertilizer was applied to winter wheat in fall at planting through

1994. Since 1995, application has occurred in early spring. Lime was applied to all
plots in September 1992 at a rate of 2T/a. Herbicides (broadcast applications) and
pesticides are applied in spring prior to planting and as needed any time thereaﬁer.
Details regarding herbicide and pesticide applications can be found on the KBS LTER
website at (http://lter.kbs.msu.edu/Data/LTER_Metadatajsp?Table=KBSO42-010).

The only external nutrient source in the organic-based plots is composted dairy
manure that is obtained from various sources, including the KBS Dairy. Compost is
applied on a dry weight basis based on estimated nitrogen, phosphorus and potassium
requirements and the rate of application takes into account moisture and nitrogen
content of the compost. Since 1995 the application rate is determined after correction
for sand content. Row cultivation and rotary hoeing is performed as necessary early in
the growing season to manage weeds in corn and soybean crops in the organic
treatments. In addition, the western half of each organic sub-plot is planted with cover
crops each year. Cover crops vary depending on the crop and include: annual ryegrass
(Lolium multiﬂorum Lam), hairy vetch (Vicia villosa Roth), red clover — Michigan
mammoth (T rifolium pratense L.), and crimson clover (T rifolium incarnatum L.).
Additional details regarding the cover crops planted in the LFL can be found at
(http://lter.kbs.msu.edu/Data/LTER Metadataisp?Table=KBSO42-007).

Biodiversity Experiment Plots

Data reported in Chapters 4 and 5 (and part of Chapter 2) were collected from
the Biodiversity Experiment Plots (BEP) at the KBS LTER. The BEP was established
in the spring (corn and soybean) and fall (winter wheat) of 2000 in an area that had

previously been used for tillage trials for the LTER experiment. The plots are located to

the west of the main LTER experiment (Figure 1.1) and are bordered by B-Avenue to
the south and 40‘h Street to the west.

The BEP was designed to determine the impacts of manipulations of row-crop
diversity on ecosystem functions related to production agriculture. The treatments
include com, soybean, and winter wheat grown at six levels of crop diversity, ranging
from continuous monocultures to two- and three-crop rotations with zero, one, and two
cover crop species annually (Table 1.1). In total, there are 21 treatments: 18 cropping
sequences (3 main crops x 6 crop-diversity treatments), two fallow treatments (spring
vs. fall) and a bare-soil (continuously plowed) treatment. Diversity treatments were
randomly applied to plots in three linear blocks; in the fourth block, treatments were
arranged in sequence in a 4 x 5 block to facilitate demonstration of treatment effects.
The four replicate blocks are separated by 8m grass buffers. Individual plots were 9.1m
x 27.4m with no buffer strips between plots (Figure 1.2).

In treatment one each crop is planted in continuous monoculture; corn was
planted in late April-early May, soybean in early May and winter wheat in October.
Following crop harvest (October and November for soybean and corn, respectively, and
July for wheat) plots were left fallow until the next season, at which time they were
tilled (chisel plowed) and the same crop planted again (Figure 1.3). Treatment two is
similar to treatment one except that a cover crop is planted each year (Figure 1.4). In
the corn and wheat monocultures red clover (T rifolium pratense L.) is sown into the
plots in July and March, respectively. Cereal rye (Secale cereale L.) is sown into
soybean stubble following harvest in October. Treatment three is a two-crop rotation

without cover crops (Figure 1.5). In one sequence soybean follows corn and in another

Table 1.1. Crop diversity treatments at the KBS LTER Biodiversity Experiment Plots
(BEP). Treatments were established in 2000 and include monocultures and two- and
three-crop rotations of corn (C), soybean (S) and wheat (W), with and without grass or
legume cover crops. NA, not applicable.

 

No. Crop species

 

 

 

Annually Sum over rotation
Treatment Sequence Rotati on/ Crops Cover Total Crops Cover Total
Group # description crops crops
BS B21 Bare-soil NA NA NA NA NA NA
1 B20 W 1 0 1 1 0 1
1 B 19 S 1 0 l 1 0 1
l B 1 8 C 1 0 1 1 0 1
2 B 1 7 W 1 1 2 1 1 2
2 B 16 S 1 1 2 1 l 2
2 B15 C l 1 2 1 1 2
3 B 14 W—S 1 0 1 2 0 2
3 B 1 3 S-C 1 0 1 2 0 2
3 B 12 C-S 1 0 1 2 0 2
4 B1 1 W-C-S 1 0 1 3 0 3
4 B 10 S-W—C l 0 l 3 0 3
4 B9 C-S-W 1 O l 3 0 3
5 BS W-C-S l 1 2 3 2 5
5 B7 S-W-C 1 1 2 3 2 5
5 B6 C-S-W 1 1 2 3 2 5
6 BS W-C-S 1 2 3 3 3 6
6 B4 S-W-C 1 2 3 3 3 6
6 B3 C-S-W 1 2 3 3 3 6
FS 132 Fallowspn-ng NA NA NA NA NA NA
FF B 1 Fallowran NA NA NA NA NA NA

 

.v x83 5 one SB No owco 5883 05 :o 9.8 m; 8—85 mo omco 5333 of
so @832 mm 5m 20:58:. .880 E0833 .5.“ 2 033. com .36 godmomom EBmBoom Such wcoq nonmam _mo_mo_o_m wmo=oM
2: a 305 EoEtomxm babe/Goa oﬁ “a 3.83 03238 So.“ E 358605 E mo om me 30:32 mEBoﬁ 92: 85 ”NA oSwE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

mm vm Nm Nm Fm ONm mm mm 9m 2m vwm mm arm 90 mm _._.m Npm 2m Om Fm nm em arm Nm 2m
mow vov mov Nov 5v ONF a: w: n: O: m: 3.— m: N: p: 05 O9 wow wow OOF mow Vow now Now 5..
um um mm mm Orm vm hm Opm mm :m ONm Om mpm 9m 2m 2m 9m 20 am Nm Om 3m mm N—m Fm
wow now wow wow 03.. ONN ON ON x. N O—N m «N 3N m—N N_.N :N OFN OON OON NON OON mON VON MON NON _.ON
mrm in arm Nrm 3m 9m 90 ONm mm Nm me Om Nrm mm 3m mm 3m 3m cm mm hm 3m rm :m mrm
m 3.. v3 m; N; :v ONO Own 9m tn O; mwm in NS” NE SN Own mom mom now com mon vow mom NON Pom

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

E . XE . ”®N_w O
Sm 2m 2m 2m 8m VVN Fm .~_n_

e. E a. a. a. ﬂoEEoEzoaxm 365235

 

 

 

 

 

 

 

Treatment 1: continuous monoculture

318‘ :1 1:1 :1 [:1 S

 

8
8
3
3 b
(D
$319 - - - - -
.E
8
g
820 .

 

1471014710147101471014710
2000 2001 2002 2003 2004

 

 

 

 

 

 

 

Month no./year

Crops

Cl corn § red clover
I soybean crimson clover
wheat cereal rye

Figure 1.3: Schematic representation of Treatment Groupl (continuous monocultures
of each crop species) in the BEP over the study period. Colored bars indicate length of
time, from crop sowing to harvest, that each species is present in each of the three
treatment sequences. Months are 1, January; 4, April; 7, July; 10, October.

10

 

Treatment 2: continuous monoculture with one cover crop annually

 

Treatment sequence

317 }

 

 

1471014710147101471014710
2000 2001 2002 2003 2004

 

 

 

 

 

 

 

Month no./year

Crops

El oom red clover
I soybean crimson clover
wheat 1:3 cereal rye

Figure 1.4: Schematic representation of Treatment Group 2 (continuous monocultures
with a cover crop) in the BEP over the study period. Colored bars indicate length of
time, from crop sowing to harvest, that each species is present in each of the three
treatment sequences. Months are 1, January; 4, April; 7, July; 10, October.

 

 

Treatment sequence

814

Treatment 3: two-crop rotation

’—

312‘ 1:3 - [:3 - S

8‘3? — :3 - :1 -

 

 

147
2000

 

10

 

1

4 7 10
2001

 

1

4 7
2002

10

 

1

4 7
2003

1014 710
2004

 

 

1:] com
I soybean

E] wheat

Crops

Month no./year

red clover

’/

é

crimson clover

cereal rye

 

Figure 1.5: Schematic representation of Treatment Group 3 (two-crop rotation) in the
BEP over the study period. Colored bars indicate length of time, from crop sowing to
harvest, that each species is present in each of the three treatment sequences. Months

are 1, January; 4, April; 7, July; 10, October.

12

corn follows soybean. In the third sequence wheat follows soybean. There is no com-
wheat sequence; therefore wheat is not present in this treatment one out of every two
years. Treatment four is a three-crop rotation (Figure 1.6). The order of crops in the
three-crop rotation is always com-soybean-wheat; however each entry point of the
rotation is present each year (i.e. three sequences: com-soybean-wheat, soybean-wheat-
com, and wheat-corn-soybean). Treatment ﬁve is similar to treatment four except that
one cover crop is planted each year (Figure 1.7). In July, red clover and crimson clover
(T rifolium incarnatum L.) are planted during the corn and soybean phases, respectively.
Red clover is frost seeded into wheat in March. Treatment six is similar to treatment
ﬁve except that an additional cover crop, cereal rye, is planted in corn and wheat
stubble in October or November (Figure 1.8). Cereal rye is not planted in the soybean
phase of the rotation. Tillage and planting times are the same for all crops across all
treatments. Data on weed abundance and composition and soil resource levels were
also collected in the spring and fall fallow treatments and are presented in the appendix
to Chapter 4. No data was collected in the bare soil treatment.

Protocols for crop management in the BEP are identical to those in the LTER
main site Treatment 4 (T4). All plots are chisel plowed and soil ﬁnished prior to
planting main crops (corn, soybean and wheat). Weeds are managed with inter-row
cultivation and rotary hoeing in corn and soybean early in the growing season. No
additional weed management is performed in wheat. Irnportantly for the questions I
was interested in addressing, there are no additions of chemicals to these plots, and

except for the prescribed differences in cover crops or crop rotations, the plots were

13

 

Treatment 4: three-crop rotation

39’ [:1 1:: -:

 

310' :1

Treatment sequence

3‘" :3 l:l

 

147101471014710‘1471014710
2000 2001 2002 2003 2004

 

 

 

 

 

 

 

Month no./year

Crops
|:l com red clover
I soybean crimson clover
E] wheat cereal rye

Figure 1.6: Schematic representation of Treatment Group 4 (three-crop rotation) in the
BEP over the study period. Colored bars indicate length of time, from crop sowing to
harvest, that each species is present in each of the three treatment sequences. Months
are 1, January; 4, April; 7, July; 10, October.

14

 

Treatment 5: three-crop rotation with one cover crop annually

86}

 

 

 

37}

 

 

 

Treatment sequence

BB

 

 

 

1471014710147101471014710
2000 2001 2002 2003 2004

 

 

 

 

 

 

 

Month no./year

Crops

1:] corn N red clover
- soybean % crimson clover
[:1 wheat cereal rye

Figure 1.7: Schematic representation of Treatment Group 5 (three-crop rotation with
one cover crop annually) in the BEP over the study period. Colored bars indicate length
of time, from crop sowing to harvest, that each species is present in each of the three
treatment sequences. Months are 1, January; 4, April; 7, July; 10, October.

15

 

 

Treatment 6: three-crop rotation with two cover crops annually

 

 

 

 

 

 

 

 

Treatment sequence

 

 

 

 

 

1471014710147101471014710
2000 2001 2002 2003 2004

 

 

 

 

 

 

 

Month no./year

Crops
E] corn red clover
I soybean crimson clover
E] wheat [:J cereal rye

Figure 1.8: Schematic representation of Treatment Group 6 (three-crop rotation with
two cover crops annually) in the BEP over the study period. Colored bars indicate
length of time, from crop sowing to harvest, that each species is present in each of the
three treatment sequences. Months are 1, January; 4, April; 7, July; 10, October.

16

managed identically. More detailed descriptions of the treatments can be found at
http://lter.kbs.msu.edu/Agronomic__Protocols/2004AgronomicProtocol%20.pdf.

The original treatment plan for the BEP called for no applications of fertilizer,
herbicide or pesticide. However, agronomic log records indicate that in 2001 fertilizer
(28% UNA solution) was applied to com (31 May) and herbicide was applied to wheat
stubble (29 August). This was apparently an accidental application. No additional
inputs have been applied to the treatments from 2002 to the present.

Questions of interest

The general question I was interested in was: Can ecological theory regarding
the role of diversity in ecosystem function be used to predict weed and crop yield
response to increasing row-crop diversity? One prediction of ecological theory is that
diverse plant communities should be more stable, in terms of community biomass,
productivity, and nutrient cycling, than simple plant communities (Cottingham et al.
2001). One way in which this can occur is when increased species richness leads to
larger differences in species’ responses to environmental variability, such that species
that respond positively compensate for those that respond negatively. However, there is
little theory to suggest how plant species should respond when a large component of
diversity is temporal (i.e. crop rotation) or how the diversity of crop species in a rotation
should impact the stability or diversity of weed species in a particular crop and if this
will differ under organic and conventional inputs. The treatments in the LF L allowed
us to examine these questions within the context of annually tilled row-crop agricultural
systems in which the number of crop species in a rotation and management inputs

varied (Chapter 3). Treatments included the extreme ends of the crop diversity

l7

gradient; simple (continuous corn, conventional management) and diverse cropping-
systems (com-com-soybean-wheat rotation, organic management). Because these
treatments had been in place since 1993 the diverse cropping-systems had completed
two cycles of the four-year rotation by the time weed communities and crop yields were
sampled in the summer of 2001.

The BEP provided an excellent system within which to test predictions related to
diversity-ecosystem function across a broad gradient in row-crop diversity. For many
ecosystem properties and processes the shape of the relationship between diversity and
ecosystem function has been shown to be positive, but saturating at higher levels of
diversity (Hooper et a1. 2005). This suggests that at the levels of diversity most
amenable to manipulation in most agricultural systems (i.e. at the low end of the
diversity gradient) small increases in crop diversity should have relatively large effects
on agroecosystem function. Again however, there is little theory to suggest how
biomass, productivity, and nutrient cycling should respond to increasing crop diversity
when a large component of diversity is temporal. I examined the response of the weed
community, crop yields, and resource availability across six levels of crop diversity
ranging from one crop species grown continuously, to two- and three-crop rotations
with zero, one, and two cover crop species annually and up to six crop species over a
three year rotation cycle. Because there were no external inputs applied to any of the
treatments in the BEP, the impacts of crop diversity on weeds, crop yields, and soil
resource levels could be analyzed without the possible confounding effects of variation

in fertilizer and herbicide inputs.

18

Organization of dissertation

The dissertation includes four empirical chapters formatted as manuscripts for
journal submission. In Chapter 2, I report the results of a study investigating the
response of the soil seed bank to manipulations of row-crop diversity (crop rotation) in
two different weed and input management systems (conventional vs. organic-based). I
also discuss the results in terms of the implications of using weed seed bank data to
infer the long-term weed suppressive potential of certain management practices. In
Chapter 3, I report the results of a study investigating how row-crop diversity and input
management affect the inter-annual variability (stability) of the emergent weed
community and crop yield. Data in these chapters were taken from treatments in the
LFL. In Chapters 4 and S, I report results from the ﬁrst three years (2002-2004) of the
BEP, a unique long-tenn row-crop diversity study that includes a broad range of crop
diversity through rotation (1-5 species over the rotation) and is conducted in the absence
of chemical inputs, which are often confounding in many other agronomic studies. In
Chapter 4, I report how increasing cropping-system diversity through manipulations of
rotations and cover crops impacts the abundance, composition and dynamics of the
emergent weed community in three crops: corn, soybean, and winter wheat. I further
examine how changes in the weed community are related to changes in soil resource
levels and light across the diversity treatments. In Chapter 5, I report how crop yields
vary across this same diversity gradient and evaluate whether these yield responses are

correlated with changes in the weed community and soil resource levels.

19

LITERATURE CITED

Altieri, MA. 1999. The ecological role of biodiversity in agroecosystems. Agriculture
Ecosystems and Environment 74: 19-31.

Chapin, F.S., Zavaleta, E.S., Eviner, V.T., Naylor, R.L., Vitousek, P.M., Reynolds,
H.L., Hooper, D.U., Lavorel, S., Sala, O.E., Hobbie, S.E., Mack, M.C., Diaz, S.
2000. Consequences of changing biodiversity. Nature 405: 234-242.

Cottingham, K.L., Brown, B.L., Lennon, J .T. 2001. Biodiversity may regulate the
temporal variability of ecological systems. Ecology Letters 4: 72-85.

Dekker, J. 1997. Weed diversity and weed management. Weed Science 45: 357-363.

Diaz, 8., Cabido, M. 2001. Vive la difference: plant functional diversity matters to
ecosystem processes. Trends in Ecology and Evolution 16: 646-655.

Doucet, C., Weaver, S.E., Hamill, A.S., Zhang, J. 1999. Separating the effects of crop
rotation from weed management on weed density and diversity. Weed Science
47, 729-735.

Drinkwater, L.E., Wagoner, P., Sarrantonio, M. 1998. Legume-based cropping systems
have reduced carbon and nitrogen losses. Nature 396: 262-265.

Fargione, J .E., Tilman, D. 2005. Diversity decreases invasion via both sampling and
complementarity effects. Ecology Letters 8: 604-611.

Hector, A., Schmid, B., Beierkuhnlein, C., Caldeira, M.C., Diemer, M.,
Dimitrakopoulos, P.G., Finn, J .A., Freitas, H., Giller, P.S., Good, J ., Hanis, R.,
Hogberg, P., Huss-Danell, K., Joshi, J ., J umpponen, A., Komer, C., Leadley,
P.W., Loreau, M., Minns, A., Mulder, C.P.H., O'Donovan, G., Otway, S.J.,
Pereira, J .S., Prinz, A., Read, D.J., Scherer-Lorenzen, M., Schulze, E.D.,
Siamantziouras, A.S.D., Spehn, E.M., Terry, A.C., Troumbis, A.Y., Woodward,
F.I., Yachi, S., Lawton, J .H. 1999. Plant diversity and productivity experiments
in European grasslands. Science 286: 1123-1127.

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., Setala, H., Symstad,
A.J., Vanderrneer, J ., Wardle, DA. 2005. Effects of biodiversity on ecosystem

functioning: A consensus of current knowledge. Ecological Monographs 75: 3-
35.

Hooper, D.U., Vitousek, PM. 1998. Effects of plant composition and diversity on
nutrient cycling. Ecological Monographs 68: 121-149.

20

Huston, MA. 1997. Hidden treatments in ecological experiments: Re-evaluating the
ecosystem function of biodiversity. 0ecologia 110: 449-460.

Kegode, G.O., Forcella, F., Clay, S. 1999. Inﬂuence of crop rotation, tillage, and
management inputs on weed seed production. Weed Science 47: 175-183.

Lambers, J .H.R., Harpole, W.S., Tilman, D., Knops, J ., Reich, PB. 2004. Mechanisms
responsible for the positive diversity-productivity relationship in Minnesota
grasslands. Ecology Letters 7: 661-668.

Levine, J .M., D'Antonio, GM. 19999. Elton revisited: a review of evidence linking
diversity and invisibility. Oikos 87: 15-26.

Liebman, M., Dyck, E. 1993. Crop rotation and intercropping strategies for weed
management. Ecological Applications 3: 92-122.

Liebman, M., Gallandt, ER. 1997. Ecological management of crop-weed interactions.
In: Ecology in Agriculture. LE Jackson, ed. San Diego, CA, USA. pp 291-343.

Liebman, M., Staver, GP. 2001. Crop diversiﬁcation for weed management. In:
Ecological Management of Agricultural Weeds. M Liebman, M Mohler, CL
Staver, eds. Cambridge University Press, Cambridge, UK. pp 322-374.

Loreau, M., Hector, A. 2001. Partitioning selection and complementarity in biodiversity
experiments. Nature 412: 72-76.

Minns, A., Finn, J ., Hector, A., Caldeira, M., Joshi, J ., Palmborg, C., Schmid, B.,
Scherer-Lorenzen, M., Spehn, E., Troumbis, A. 2001. The functioning of
European grassland ecosystems: potential beneﬁts of biodiversity to agriculture.
Outlook on Agriculture 30: 179-185.

Naeem, S., Knops, J .M.H., Tilman, D., Howe, K.M., Kennedy, T., Gale, S. 2000. Plant
diversity increases resistance to invasion in the absence of covarying extrinsic
factors. Oikos 91: 97-108.

Reganold, J .P., Glover, J .D., Andrews, P.K., Hinman, HR. 2001. Sustainability of three
apple production systems. Nature 410: 926-930.

Robertson, G.P., Klingensmith, K.M., Klug, M.J., Paul, E.A., Crum, J .R., Ellis, B.G.
1997. Soil resources, microbial activity, and primary production across an
agricultural ecosystem. Ecological Applications 7: 158-170.

Robertson, G.P., Swinton, SM. 2005. Reconciling agricultural productivity and

environmental integrity: a grand challenge for agriculture. Frontiers in Ecology
and the Enviornment 3: 38-46.

21

Sanchez, J .E., Harwood, R.R., Willson, T.C., Kizilkaya, K., Smeenk, J ., Parker, E.,
Paul, E.A., Knezek, B.D., Robertson, GP. 2004. Managing soil carbon and

nitrogen for productivity and environmental quality. Agronomy Journal 96: 769-
775.

Sanchez, J .E., Willson, T.C., Kizilkaya, K., Parker, E., Harwood, RR. 2001. Enhancing
the mineralizable nitrogen pool through substrate diversity in long term cropping
systems. Soil Science Society of America Journal 65: 1442-1447.

Schlapfer, F., Schmid, B. 1999. Ecosystem effects of biodiversity: A classiﬁcation of

hypotheses and exploration of empirical results. Ecological Applications 9: 893-
912.

Spehn, E.M., Hector, A., Joshi, J ., Scherer-Lorenzen, M., Schmid, B., Bazeley-White,
E., Beierkuhnlein, C., Caldeira, M.C., Diemer, M., Dimitrakopoulos, P.G., Finn,
J .A., Freitas, H., Giller, P.S., Good, J ., Hanis, R., Hogberg, P., Huss-Danell, K.,
Jumpponen, A., Koricheva, J ., Leadley, P.W., Loreau, M., Minns, A., Mulder,
C.P.H., O'Donovan, G., Otway, S.J., Palmborg, C., Pereira, J.S., Pﬁsterer, A.B.,
Prinz, A., Read, D.J., Schulze, E.D., Siamantziouras, A.S.D., Terry, A.C.,
Troumbis, A.Y., Woodward, F.I., Yachi, S., Lawton, J .H. 2005. Ecosystem
effects of biodiversity manipulations in European grasslands. Ecological
Monographs 75: 37-63.

Teasdale, J .R., Mangum, R.W., Radhakrishnan, J ., Cavigelli, MA. 2004. Weed
seedbank dynamics in three organic farming crop rotations. Agronomy Journal
96: 1429-1435.

Tilman, D., Cassman, K.G., Matson, P.A., Naylor, R., Polasky, S. 2002. Agricultural
sustainability and intensive production practices. Nature 418: 671-677.

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.

Vandermeer, J. 1989. The ecology of intercropping. Cambridge University Press,
Cambridge.

Wardle, DA. 1999. Is "sampling effect" a problem for experiments investigating
biodiversity-ecosystem function relationships? Oikos 87: 403-407.

Westerman, P.R., Liebman, M., Menalled, F .D., Heggenstaller, A.H., Hartzler, R.G.,
Dixon, PM. 2005. Are many little hammers effective? Velvetleaf (Abutilon

theophrasti) population dynamics in two- and four-year crop rotation systems.
Weed Science 53: 382-392.

Zhu, Y.Y., Chen, H.R., Fan, J.H., Wang, Y.Y., Li, Y., Chen, J.B., Fan, J.X., Yang, 8.8.,

22

Hu, L.P., Leung, H., Mew, T.W., Teng, P.S., Wang, Z.H., Mundt, CC. 2000.
Genetic diversity and disease control in rice. Nature 406: 718-722.

23

CHAPTER 2
EVIDENCE OF RAPID SEED BANK DYNAMICS IN ROW CROPS
ABSTRACT

The dynamics of soil seed banks in crop rotations of maize, soybean, and winter
wheat were investigated to determine whether weed seed inputs associated with earlier
phases of the rotation persist in subsequent phases. In one study we compared seed
banks in maize grown continuously and in a long-term rotation that included winter
wheat. A second study followed seed bank composition and abundance over three years
in plots that were planted to maize and soybean in successive years following planting
to winter wheat in 2001. Seed banks were measured by direct germination in a heated
glasshouse. Seed banks in continuous maize and the maize phase of the rotation did not
differ in species composition or abundance despite signiﬁcant differences in the winter
wheat phase of the rotation. Seed bank abundance and composition in plots planted to
winter wheat in 2001 quickly changed following planting of maize and soybean in 2002
and 2003, respectively. Data from the two experiments suggest that seed banks in
annual row cr0ps are highly dynamic and are strongly inﬂuenced by the most recent
crop. This may limit our ability to infer longer-term effects of weed management
practices from studies of soil seed banks.

INTRODUCTION

Seed banks in annual row-crops are assumed to contain seeds that have been
shed over multiple growing seasons, reﬂecting the sum effects of past and present weed
management on weed survival and reproduction (Cavers and Benoit 1989, Clements et

al. 1996, Dorado et al. 1999). The assumption that seed banks are a memory of past

24

management is a long-standing paradigm in weed ecology (Cavers and Benoit 1989,
Swanton and Booth 2004) and has often been invoked as justiﬁcation for using the seed
bank to measure the long-term weed suppression potential of a given management
system (Cardina et al. 2002, Moonen and Barberi 2004, Swanton and Booth 2004).
However, this assumption may not be valid if weed seed banks do not persist for more
than a few growing seasons or are highly dynamic.

The ability of seeds to persist in the soil depends on a number of physiological,
environmental, and evolutionary factors. In desert systems, for instance, annual plant
species have evolved dormancy as a bet-hedging strategy to maintain populations in
environments that are temporally variable, and a large percentage of the seed bank
remains dormant even when conditions are conducive to germination (Philippi 1993,
Pake and Venable 1996). In contrast, most agricultural weeds have evolved to
maximize growth and reproduction during favorable conditions (Grime 1977) and
therefore germinate when the appropriate micro-environmental cues are detected.
Although there are a few well-known examples of seeds of some arable weed species
persisting for relatively long periods of time and under controlled conditions (Cavers
and Benoit 1989, Telewski and Zeevaart 2002), the majority of studies in cultivated
systems suggest that most seeds germinate or die in the ﬁrst few years after entering the
soil seed bank (Cavers and Benoit 1989, Lutrnan et al. 2002, Teasdale et al. 2004). The
discrepancy between such studies and assumptions of long-tenn seed bank persistence
indicates that the dynamics of seed bank communities in annual row crop systems are

still not well understood.

25

Crop rotations provide useful systems for evaluating the persistence and
dynamics of weed seed banks because weed management, and therefore seed inputs to
the soil, varies with each crop in rotation (Liebman and Dyck 1993). Because certain
crops are often associated with distinct weed communities (Cavers and Benoit 1989,
Squire et al. 2000), inputs of seed during that phase of the rotation can leave a unique
signature within the seed bank. The persistence of this seed input signature can be
measured in successive crop phases to determine how long seed inputs remain in the
seed bank.

The aim of this study was to examine weed seed bank dynamics and determine
the relative persistence of seed inputs to the soil in annually tilled row crops.
Speciﬁcally, we were interested in whether weed seed inputs associated with the winter
wheat (T riticum aestivum L.) phase of a rotation were apparent in the seed bank in
' successive phases of a maize (Zea mays L.)-soybean (Glycine max L.)-winter wheat
rotation. F all-sown crops such as winter wheat have been shown to be associated with a
weed ﬂora that is distinct from spring-sown crops (Squire et a1. 2000, Crawley 2004)
such as maize and soybean. Data from two studies were examined to investigate seed
bank dynamics and persistence. The ﬁrst study compared weed seed banks in maize
grown continuously and in a long-term crop rotation with soybean and winter wheat
under conventional and organic-based input management. The two input management
systems were compared to determine whether the method of weed control and nutrient
inputs affected seed bank persistence. The second study followed seed bank abundance
and composition in replicate plots that were planted to winter wheat in 2001, maize in

2002 and soybean in 2003 in 'a rotation experiment managed without external chemical

26

inputs. Seed banks were sampled each year in spring and the composition and
abundance of seeds in the soil assessed by direct germination in a heated glasshouse.
MATERIALS AND METHODS
Study Sites

Two studies were carried out in experimental plots located at the W. K. Kellogg
Biological Station Long Term Ecological Research (KBS LTER) site in SW Michigan,
USA. Soils at both study sites were a mixture of Kalamazoo (ﬁne-loamy, mixed, mesic
Typic Hapludalfs) and Oshtemo (coarse—loamy, mixed, mesic Typic Hapludalfs) sandy
loams (Sanchez et a1. 2001). Annual precipitation is 90 cm; about half as snow, and
mean annual temperature is 9.70 C. Detailed site and soil and management descriptions

of the site are available at http://lter.kbs.msu.edu.

 

The ﬁrst study was conducted in 2001 at the Living Field Lab experimental
plots (hereafter LF L). The LFL was established in 1993 to examine the effects of
extended rotations and different input management systems on nutrient cycling and
agroecosystem function. The site was planted to Income (Medicago sativa L.) in 1992
and rotation treatments consisting of continuous maize and each phase of a four-year
maize - maize -soybean-winter wheat rotation were initiated in 1993. Rotation
treatments were randomly assigned to plots within blocks that received one of four
input management treatments in a split-plot design. All phases of the rotation were
present each year. For this study we focused only on continuous maize (CM) and the
second year maize (rotation maize, RM) and winter wheat (rotation wheat, RW) phases
of the rotation in two management systems; a conventional system that received full

inputs of fertilizer, herbicides (broadcast applications) and pesticides, at rates

27

recommended for the region (Sanchez et al. 2004), and an organic-based system which
received composted dairy manure, but no chemical inputs. Weed management in the
organic input system was via periodic (2-4 times/yr) inter-row cultivation (row
cultivation and rotary hoeing). Each phase represented different times since the last
appearance of winter wheat in the rotation. Plots planted to continuous maize (CM)
had never been planted to winter wheat; second-year maize plots (RM) had been
planted to maize the previous year and to winter wheat three and seven years prior to
sampling; and winter wheat plots (RW) had been planted to wheat the previous season
and ﬁve years prior. Plots in both input systems were chisel plowed and soil ﬁnished
each year prior to planting. All treatments (management system x rotation) were
replicated four times. Plot sizes differed between the two input treatments and
measured 15 x 9 m (conventional) and 15 x 4.5 m (organic).

The second study was conducted from 2002-2004 at the Biodiversity
Experiment Plots (hereafter BEP) approximately one km south of the LF L site. The
BEP was established in 2000 to investigate the effects of manipulating row-crop
diversity on ecosystem ﬁmctioning. Treatments in the BEP include continuous
monocultures and two and three-crop rotations of maize, soybean, and winter wheat
grown with and without cover crops. The BEP was managed without external chemical
inputs and plots were chisel plowed and soil ﬁnished each year prior to planting. All
phases of the rotation treatments were present each year. Plots were 9 x 27 m and
treatments were arranged in a randomized complete block with four replications. For

this study we followed seed bank abundance and composition in four replicate plots that

28

were planted to winter wheat in 2001, maize in 2002, and soybean in 2003 and
compared these to the seed banks of continuous monocultures of the same crops.
Seed bank sampling

Soil seed banks were sampled in a similar manner in both the LF L and BEP
studies. In 2001 (LF L study) and 2002-2004 (BEP study) soil seed banks were sampled
by taking ten soil cores (2 cm diameter) to a depth of 5 cm from a of 25 x 25 cm area in
three central locations in each plot. Sampling occurred in mid-May after the plots had
been tilled and crops planted, but prior to crop emergence. Sampling in the spring has
been shown to provide reliable estimates of the viable seed bank (F orcella 1992)
because it allows natural dormancy-breaking mechanisms to operate over the winter.
Sampling immediately after tillage and planting also ensured that seed banks were
thoroughly mixed. Each ten-core soil sample was composited and spread on sterile
growing medium (Sunshine Germinating Mix #3, Sun Gro Horticulture, Bellevue,
Washington, USA) in 25 x 25 cm half-ﬂats in a glasshouse. Flats were watered daily
from above with a mist sprayer to keep the soil surface moist. Flats were monitored for
germination approximately twice a week from May to October. Seedlings that emerged
from the ﬂats were identiﬁed and removed. Seedlings that could not be identiﬁed
immediately were transplanted and grown in separate pots until identiﬁcation was
possible.

Statistical analyses

For each soil sample collected from the LF L, we calculated the density (total

seedlings emerged) and species richness (number of species). Data were averaged for

the three samples taken from each plot, and plot means used as the unit of replication.

29

The effects of management system and rotation on seed bank density and species
richness were analyzed using a split-plot design with management system
(conventional, organic) as the whole plot factor and rotation phase (continuous maize,
rotation maize, rotation wheat) as the subplot factor. Analyses of variance were
performed using the MIXED procedure in SAS (SAS Version 8.02; SAS Institute, Cary,
NC, USA). Treatment means were separated with errors appropriate for a split-plot
design using the Tukey-Kramer Test at the P = 0.05 level. Seedling density and
species richness data were square root transformed prior to analysis to satisfy the
assumptions of AN OVA. Seedling density data are expressed as number of emerged
weed seedlings per m2 (assuming 25 cm2 per sample). Data were back transformed for
presentation in tables and ﬁgures.

To visualize plot-level patterns in species abundance and composition among
the rotation and management system treatments, we performed non-metric
multidimensional scaling, (NMDS: Mather 1976, McCune and Grace 2002) using the
PC-ORD package (version 4.25, McCune and Mefford 1999). NMDS is an ordination
method that maximizes rank order correlation between distances derived from the
original dataset and those in ordination space, and is often the most appropriate
ordination method for community datasets (McCune and Grace 2002). Prior to
ordination, seedling abundance values were log (x + 1) transformed and a distance
matrix was calculated using the Bray-Curtis distance metric. Forty runs of the
ordination (at random starting conﬁgurations and with a maximum of 400 iterations per

run) were performed using an instability criterion of 0.00001. These runs were

30

compared to 50 randomized runs to assess the signiﬁcance of the reduction in stress
from six to one dimensions (Monte-Carlo Test).

In addition to the ordination, we used non-parametric distance-based
multivariate analysis (DISTLM v.5, Anderson 2001, 2004) to test for signiﬁcant
differences in seed bank community composition and abundance among the treatments.
Main effects of inputs, rotation phase, and their interaction on species composition and
abundance were tested by permutation (9999 permutations) using error terms
appropriate for a split plot design. The test statistic (pseudo F) is analogous to Fisher’s
F-ratio and was calculated from a distance matrix of Bray-Curtis dissimilarity
coefﬁcients. Species abundance values were log (x +1) transformed prior to analysis.

Seed bank data from the BEP were analyzed using NMDS ordination to
determine how seed bank composition and abundance changed when plots planted to
winter wheat in 2001 were subsequently planted to maize in 2002 and soybean in 2003.
NMDS was performed on the species abundance values measured for the four replicate
plots of the rotation treatment each year from 2002-2004 and the four replicates of the
continuous monoculture that corresponded to the crop planted in the rotation replicates
(i.e. continuous monoculture of winter wheat in 2001, maize in 2002, and soybean in
2003). Seedling abundance data were log (x +1) transformed prior to calculating Bray-
Curtis distances and performing ordination.

RESULTS
LFL Seed bank density and species richness
In the LF L plots, there was no effect of input management system (organic vs.

conventional) on seed bank density or species richness (density: P = 0.76; richness: P =

31

0.14); however, there was a signiﬁcant effect of rotation phase (density: F 2,. 2 = 4.89 , P
= 0.03 ; richness: F2,” = , P = 0.007). In both input systems, seed bank density and
species richness were signiﬁcantly greater in the winter wheat phase (RW, last planted
to winter wheat in 2000) than the rotation maize phase (RM, last planted to winter
wheat in 1998) or continuous maize monoculture (CM) (Figure 2.1). Seed bank density
and richness did not differ between the two maize treatments (RM and CM); and there
was no interaction between input management system and rotation phase on seed bank
density or richness (density: P = 0.99; richness: P = 0.26).
LFL Community abundance and composition

Thirty-seven species were identiﬁed from the soil samples collected in the LF L
(Table 2.1). Of these, six were unique to the wheat phase of the rotation (RW). These
tended to be monocots that were present at relatively low relative densities within the
seed bank (~ < 1%).

The NMDS ordination reduced the dimensionality of the LFL seed bank dataset
to three main axes (minimum stress = 9.86, P = 0.02), which accounted for 58.2, 19.7,
and 14.0 % (cumulative r2 = 91.9%) of the information in the analytical data set,
respectively. Ordination showed a clear distinction in seed bank abundance and
composition between the winter wheat phase of the rotation and the continuous and
rotation maize (Figure 2.2). The non-parametric multivariate analysis conﬁrmed that
differences among seed banks were due to the effect of rotation phase (rotation phase:
pseudo F2,” = 8.82, P = 0.0001) and that effects of input management and their

interaction were not signiﬁcant (inputs: P = 0.055; interaction: P = 0.084).

32

(A) (B)

 

 

 

 

 

 

 

 

.— 14.—
1400 b b
1200~ 1,- 12—
a.“ o.
E 1000— E 10—
7,: 2.
on
E 800'" E 8”
B c
8 5
a soar -: 6-
3’s 8
“g 400* g 4—
L” a)
200- 2—
O ___l

 

Figure 2.1: Density (A) and species richness (B) in May 2001 of seedlings emerged
from the soil seed bank in continuous maize (CM, never planted to winter wheat) and
the rotation phase of maize (RM, planted to winter wheat in 1998) and winter wheat
(RW, planted to winter wheat in 2000) of a four year maize-maize-soybean-winter
wheat rotation under conventional (black bars) and organic-based (white bars) nutrient
and weed management. Bars are means :1: 1 SE. There was no effect of input system
(organic vs. conventional) so bars sharing the same lower case letter indicate no
signiﬁcant difference among rotation phases at P = 0.05 (Tukey-Kramer Test).

33

Table 2.1: Relative density (%) of weed species that emerged from the soil seed bank
samples collected in May 2001 in continuous maize (CM, never planted to winter
wheat) and the rotational phase of maize (RM, last planted to wheat in 1998) and winter
wheat (RW, last planted to wheat in 2000) phases of the rotation.

 

Rotation phase

 

 

Weed species CM RM RW
Amaranthus retroflexus L. 0.00 0.59 1.03
Ambrosia artemisiifo/ia L. 0.00 0.24 1.03
Arabidopsis thaliana (L.) Heynh. 9.97 8.83 3.59
Aster sp. 0.00 0.12 0.00
Capse/Ia bursa—pastoris (L) Medicus 0.00 1.18 0.00
Cardamine hirsuta L. 0.26 0.00 0.00
Chenopodium album L. 7.93 11.43 1.76
Conyza canadensis (L.) Cronq. 0.00 0.12 0.59
Digitaria ischaemum (Schreb. ex Schweig.) Schreb 0.13 0.71 1.54
Digitaria sanguine/is (L.) Scop. 0.90 0.47 3.45
Echinochloa crus-galli (L.) Beauv. 0.26 0.00 0.00
Eleusine indica (L.) Gaertn. 0.38 0.12 0.00
Epilobium sp. 0.00 0.00 0.07
Eragrostis cilianensis (AIL) E.Mosher 0.00 0.00 0.07
En'geron strigosus Muhl. ex Willd. 0.64 0.00 0.07
Juncus tenuis Willd. 0.13 0.00 0.07
Lepidium virginicum L. 0.13 0.12 0.00
Mollugo Verticillata L. 0.26 14.72 0.07
Oxalis stricta L. 1.15 2.36 4.91
Panicum capillare L. 0.00 0.00 0.07
Panicum dichotomiﬂorum Michx. 0.00 0.24 5.94
Panicum sp. 0.00 0.00 0.15
Plantago major L. 0.00 0.35 3.96
Poa sp. 11.89 3.06 0.51
Polygonum aviculare L. 0.00 0.00 0.29
Polygonum convolvulus L. 0.26 0.24 0.00
Polygonum persican’a L. 0.13 1.53 1.91
Portulaca oleracea L. 0.64 1.18 0.51
Setaria faberi Herrm. 0.00 0.00 0.07
Setaria glauca (L) Beauv. 0.00 0.12 0.07
Setaria viridis (L.) Beauv. 0.00 0.12 0.07
Solanum ptycanthum Dun. 0.00 0.12 0.15
Ste/[aria media (L) Vill. 22.63 23.79 2.49
Taraxacum oﬁicinale Weber in Wiggers 11.51 5.18 54.62
Trifolium repens L. 0.13 0.00 0.07
Veronica peregrine L. 29.80 22.03 10.56
Veronica persica Poir. 0.90 1.06 0.29

 

34

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Axrs 2
4

 

 

 

 

 

 

 

Axis 1

Figure 2.2: NMDS ordination of weed seed banks from soils collected in May 2001 in
continuous maize (never planted to winter wheat) (triangles) and the rotation phase of
maize (last planted to winter wheat in 1998) (inverted triangles) and winter wheat (last
planted to winter wheat in 2000) (squares) of a four year maize-maize-soybean-winter
wheat rotation under conventional (ﬁlled symbols) and organic-based (open symbols)
nutrient and weed management. Values plotted are replicate scores. Only the two most
explanatory axes (r2 = 0.779) are displayed.

35

BEP Seed bank dynamics

In the BEP plots the seed bank was measured each year in four replicate rotation
plots, and the corresponding continuous monocultures. Eight weed species were unique
to both the continuous wheat and rotation plots planted to wheat in 2001 (Table 2.2).
After two years, following planting to maize (2002) and soybean (2003), only three of
these species were still detectable in the seed bank of the rotation replicates in 2004.
The species still present in 2004 tended also to be associated with maize (Table 2.1).

The NMDS ordination performed on the BEP species abundance data reduced
dimensionality to three main axes (minimum stress = 11.30, P = 0.020), which
explained 49.6, 28.3, and 8.7% (cumulative r2 = 86.6%) of the information in the
analytical dataset. The ordination showed that seed banks in the continuous
monocultures of wheat, maize, and soybean were distinct from one another, and that
there were dramatic changes in seed bank abundance and composition from year to year
in the crop rotation replicates (Figure 2.3). Following planting to wheat in 2001, seed
bank abundance and composition in three of the four rotation replicates was similar to
that of continuous wheat in 2002. In 2003, after a year of maize, seed bank abundance
and composition had shifted towards that of continuous maize. In 2004, the abundance
and composition of the seed bank had again shifted, this time in the direction of
continuous soybean. The seed bank in the fourth replicate was less similar to that of
continuous wheat in 2002, but still showed a similar pattern of change following

planting to maize and soybean.

36

Table 2.2: Relative densities (%) in the soil seed bank of the eight weed species
common to both the continuous winter wheat monoculture and the wheat rotation in
2001 and following planting of maize (2002) and soybean (2003).

 

 

Weed species 2001 2002 2003
Wheat Maize Soybean
Chenopodium album L. 4.8 37.1 25.5
Digitaria ischaemum (Schreb. ex Schweig.) 1.6 0 0
Schreb
Digitaria sanguinalis (L.) Scop. 8.0 2.9 0
Erigeron spp. 1.6 0 0
Mollugo verticillata L. 34.9 2.9 9.6
Panicum dichotomiﬂorum Michx. 3.2 0 0
Potentilla norvegica L. 3.2 0 0
Stellaria media (L.) Vill. 25.4 51.4 46.8

 

37

,1 wheat

V [V2002

 

2009/

I
, v. ---------------- " ‘9
__~
I’ ’1’ - ”I" 2002
Axis1 ,. maize
. / ﬂ " V
l ’ I
1’ ,’ V ;”T I"
/’ ,,’ I I '
1’ ,I r”’ Ii
[I ’I ’I’ z
1’ ,1 ’fft T'T I 2004
a I

’5’ .-:——r‘ 2003 5~.

It— 2004
2003 $2004

soybean

 

AAA

Figure 2.3: NMDS ordination of weed seed banks in four replicate plots (ﬁlled
symbols) in 2001, 2002, and 2003 following planting to winter wheat (inverted
triangles), maize (squares), and soybean (triangle), respectively, and the corresponding
continuous monoculture in that year. Replicate plots are labeled 1-4 and dashed arrows
indicate each plot’s trajectory through time. Ellipses around monocultures (open
symbols) are for illustrative purposes and do not imply conﬁdence intervals. Only the
two most explanatory axes (r2 = 0.780) are displayed.

38

DISCUSSION

Results of both the LF L and BEP studies showed that weed seed banks in the
crop rotations were highly dynamic. Other studies have found similar rapid changes in
the abundance and composition of the seed bank from year to year (Leck and Leck
1998, Teasdale et a1. 2004). For example, Buhler et al. (2001) found that the seed bank
changed with each crop over a ﬁve-year rotation that included hay, maize, soyabean,
and cat; and that abundances of seeds of Amaranthus rudis Sauer (common waterhemp)
and Setaria species (foxtail) changed 2 to 33-fold from crop to crop. Seed bank density
in an oldﬁeld increased 14-fold in just two years following abandonment from
agriculture (Leek and Lock 1998).

The rapid dynamics observed in these studies were likely due to two factors: low
seed bank persistence and large differences in the abundance and composition of seed
inputs associated with each crop. We found little evidence that seed banks were highly
persistent in either the LF L or BEP studies, or that persistence was affected by the
system of nutrient and weed management. Other researchers have found a similar lack
of long—term persistence of seed banks in agricultural ﬁeld studies. A summary of crop
management studies found that most viable seeds were lost from the seed bank
following one to four years of reduced or eliminated seed inputs (Cavers and Benoit
1989). Buhler (1999) found that Setariafaberi Herrm. (giant foxtail) seed densities in
the soil decreased rapidly when seed inputs were eliminated. In a controlled ﬁeld study,
Lutman et al. (2002) reported annual seed loss rates from 20 to 58% for sixteen weed

species.

39

The observation from the LF L study that a few species were present in both the
winter wheat and maize rotation phases, but not continuous maize (Table 2.1) suggests
that there is limited short-tenn persistence in the seed bank. Studies reporting higher
species diversity in crop rotations compared to continuous monocultures (Liebman and
Dyck 1993, Dorado et al. 1999, Cardina et a1. 2002) support the notion of limited short-
tenn persistence in some species. The few seeds in the soil that do persist for more than
one or two seasons are likely important for the maintenance of populations across
rotation cycles (Liebman and Dyck 1993, Buhler and Hartzler 2001). However,
because these persistent seeds are present at such low population levels, their detection
within the seed bank may be difﬁcult and may not provide useful information regarding
past weed management. While external factors such as seed predators or other biotic
agents (Brust and House 1988, Liebman and Dyck 1993) could have affected seed bank
persistence, there is little a priori reason to expect long-term seed bank persistence in
most arable systems, as most weed species have evolved in environments that are
conducive to rapid growth and reproduction rather than prolonged dormancy (Grime
1977, Clements et al. 2004).

Annual inputs of seed to the soil are a consequence of “ﬁlters” operating within
the context of the cropping system (Cavers and Benoit 1989, Booth and Swanton 2002).
These ﬁlters exclude some species and allow others to establish and reproduce, and can
include the crop and timing of crop sowing (Hald 1999, Hallgren et al. 1999, Menalled
et al. 2001, Booth and Swanton, 2002, Crawley 2004). For example, a comparison of
the emergent weed communities under three different 6-yr crop rotations in southern

Alberta found the species of crop was a much more important factor inﬂuencing the

40

weed ﬂora than crop rotation (Andersson and Milberg 1998). Similarly, Doucet et al.
(1999) found that when continuous monocultures and 3-crop rotations containing
maize, soybean, and winter wheat that had been in place for nine years were planted to
maize in the tenth year there were few residual effects of the rotation on the emergent
weed community. In our study, the strong ﬁltering effects of different crops, coupled
with the lack of long-term seed bank persistence, likely contributed to the observed
similarities in seed banks between crops in rotation and monocultures of the same crop.

The relatively rapid seed bank dynamics and low degree of seed bank
persistence observed in the present study are not likely speciﬁc to crop rotations. Crop
rotations only make these phenomena more apparent. Given the rapid dynamics and
short persistence times of arable weed seeds in the soil, we suggest caution be exercised
when using the seed bank to infer past trends in weed control or weed seed production,
or evaluate the weed management efﬁcacy of a given cropping system. Seed banks at
any given point in time should likely be viewed more as a reﬂection of the weed
management success or failures that occurred the previous season rather than the
cumulative effects of management occurring over longer-time scales. Therefore, it may
be unwise to use the seed bank to infer the long-term weed suppression potential of a
given management system or to relate the size and composition of the seed bank to
weed management practices occurring over time scales longer than one or two cropping
seasons.

ACKNOWLEDGMENTS
We acknowledge and thank Dr. R. R. Harwood for allowing us to sample the

Living Field Lab (LF L) and for having the foresight to design and establish this

41

experiment. We also thank G. Parker and J. Simmons (farm managers, LF L and BEP,
respectively) and J. Smeenk for their work in maintaining these treatments. A. Smith
and M. Hammond assisted with ﬁeld sampling and greenhouse work. D. Buhler, S.
Emery, E. Garcia, D. Landis, and W. Mahaney provided helpful comments on earlier
drafts of this chapter. Support for this study was provided by the NSF -ﬁinded KBS
LTER project (DEB98-10220; KLG) and by a fellowship from the CS. Mott Program

in Sustainable Agriculture (RGS).

42

LITERATURE CITED

Anderson, M.J. 2001. A new method for non-parametric multivariate analysis of
variance. Austral Ecology 26: 32-46.

Anderson, M.J. 2004. DISTLM v.5: a FORTRAN computer program to calculate a
distance-based multivariate analysis for a linear model. Department of Statistics,
University of Aukland, New Zealand.

Andersson, T.N., Milberg, P. 1998. Weed ﬂora and the relative importance of site, crop,
crop rotation, and nitrogen. Weed Science 46: 30-38.

Booth, B.D., Swanton, C.J. 2002. Assembly theory applied to weed communities. Weed
Science 50: 2-13.

Brust, G.E., House, G.J. 1988. Weed seed destruction by arthropods and rodents in low-
input soybean agroecosystems. American Journal of Alternative Agriculture 3:
19-25.

Buhler, DD. 1999. Weed population responses to weed control practices. 1. Seed bank,
weed populations, and crop yields. Weed Science 47: 416-422.

Buhler, D.D., Hartzler, R.G. 2001. Emergence and persistence of seed of velvetleaf,
common waterhemp, woolly cupgrass, and giant foxtail. Weed Science 49: 230-
235.

Buhler, D.D., Kohler, K.A., Thompson, R.L. 2001. Weed seed bank dynamics during a
ﬁve-year crop rotation. Weed Technology 15: 170-176.

Cardina, J ., Henns, C.P., Doohan, DJ. 2002. Crop rotation and tillage system effects on
weed seedbanks. Weed Science 50: 448-460.

Cavers, P.B., Benoit, D.L. 1989. Seed banks in arable land. In: Ecology of Soil Seed
Banks, (eds. M.A. Leck, V.T. Parker, R.L. Simpson), 309-328. Academic Press,
San Diego, USA.

Clements, D.R., Benoit, D.L., Murphy, S.D., Swanton, C]. 1996. Tillage effects on
weed seed return and seedbank composition. Weed Science 44: 314-322.

Clements, D.R., DiTommaso, A., Jordan, N., Booth, B.D., Cardina, J ., Doohan, D.,
Mohler, C.L., Murphy, S.D., Swanton, C.J. 2004. Adaptability of plants
invading north American cropland. Agriculture, Ecosystems and Environment
104: 379-398.

Crawley, M.J. 2004. Timing of disturbance and coexistence in a species-rich ruderal
plant community. Ecology 85: 3277—3288.

43

Dorado, J ., Del Monte, J .P., Lopez-Fando, C. 1999. Weed seedbank response to crop
rotation and tillage in semiarid agroecosystems. Weed Science 47: 67-73.

Doucet, C., Weaver, S.E., Hamill, A.S., Zhang, J. 1999. Separating the effects of crop
rotation from weed management on weed density and diversity. Weed Science
47: 729-735.

F orcella, F. 1992. Prediction of weed seedling densities from buried seed reserves.
Weed Research 32: 29-38.

Grime, J .P. 1977. Evidence for existence of 3 primary strategies in plants and its
relevance to ecological and evolutionary theory. American Naturalist 111: 1169-
1 194.

Hald, AB. 1999. The impact of changing the season in which cereals are sown on the
diversity of the weed ﬂora in rotational ﬁelds in Denmark. Journal of Applied
Ecology 36: 24-32.

Hallgren, E., Palmer, M.W., Milberg, P. 1999. Data diving with cross-validation: an
investigation of broad-scale gradients in Swedish weed communities. Journal of
Ecology 87: 1037-1051.

Leck, M.A., Leck, CF. 1998. A ten-year seed bank study of old ﬁeld succession in
central New Jersey. Journal of the Torrey Botanical Society 125: 11-32.

Liebman, M., Dyck, E. 1993. Crop rotation and intercropping strategies for weed
management. Ecological Applications 3: 92-122.

Lutman, P.J.W., Cussans, G.W., Wright, K.J., Wilson, B.J., Wright, G.M., Lawson,
HM. 2002. The persistence of seeds of 16 weed species over six years in two
arable ﬁelds. Weed Research 42: 231-241.

Mather, RM. 1976. Computational Methods of Multivariate Analysis in Physical
Geography. J. Wiley & Sons, London, UK.

McCune, B., Grace, J .B. 2002. Analysis of Ecological Communities. MjM Software
Design, Gleneden Beach, USA.

McCune, B., Mefford, M.J. 1999. PC-ORD. Multivariate Analysis of Ecological Data,
Version 4. MjM software Design, Gleneden Beach, Oregon, USA.

Menalled, F.D., Gross, K.L., Hammond, M. 2001. Weed aboveground and seedbank

community responses to agricultural management systems. Ecological
Applications 11: 1586-1601.

44

Moonen, A.C., Barberi, P. 2004. Size and composition of the weed seedbank after 7
years of different cover-crop-maize management systems. Weed Research 44:
163-177.

Pake, C.E., Venable, D.L. 1996. Seed banks in desert annuals: Implications for
persistence and coexistence in variable environments. Ecology 77: 1427-1435.

Philippi, T. 1993. Bet-hedging germination of desert annuals-beyond the lst year.
American Naturalist 142: 474-487.

Sanchez, J .E., Willson, T.C., Kizilkaya, K., Parker, E., Harwood, RR. 2001. Enhancing
the mineralizable nitrogen pool through substrate diversity in long term cropping
systems. Soil Science Society of America Journal 65: 1442-1447.

Sanchez, J .E., Harwood, R.R., Willson, T.C., Kizilkaya, K., Smeenk, J ., Parker, E.,
Paul, E.A., Knezek, B.D., Robertson, GP. 2004. Managing soil carbon and
nitrogen for productivity and environmental quality. Agronomy Journal 96: 769-
775.

Squire, G.R., Rodger, S., Wright, G. 2000. Community-scale seedbank response to less
intense rotation and reduced herbicide input at three sites. Annals of Applied
Biology 136: 47-57.

Swanton, C.J., Booth, RD. 2004. Management of weed seedbanks in the context of
populations and communities. Weed Technology 18: 1496-1502 Suppl. S

Teasdale, J .R., Mangum, R.W., Radhakrishnan, J ., Cavigelli, MA. 2004. Weed
seedbank dynamics in three organic farming crop rotations. Agronomy Journal
96: 1429-1435.

Telewski, F .W., Zeevaart, J AD. 2002. The 120-yr period for Dr. Beal's seed viability
experiment. American Journal of Botany 89: 1285-1288.

45

CHAPTER 3

WEED COMMUNITY AND CROP YIELD VARIABILITY IN DIVERSE CORN
AGROECOSYSTEMS

ABSTRACT

The effects of crop rotation and input management on inter-annual variability in
weed communities and crop yields were assessed in a four-year study in corn in SW
Michigan, USA. Variability of the weed community and corn yields was assessed using
the coefﬁcient of variation (CV) and a multivariate dissimilarity index (Bray-Curtis)
that accounted for changes in both weed species abundance and composition. The
treatments included two rotations: continuous corn and a com-com-soybean-wheat
rotation, and two input management systems: conventional (CONV) and organic-based
(ORG). Weed biomass was signiﬁcantly higher in the ORG input system; however,
there was no effect of crop rotation on weed biomass or number of weed species in a
treatment (species richness). Inter-annual variability in weed community composition
and structure was affected by both crop rotation and input management system and was
highest in the ORG rotation. In contrast to the weed community, variability in corn
yield was highest in the least diverse cropping system (CONV monoculture), despite
that system having a more constant weed community. Corn yield in the ORG rotation
was not signiﬁcantly different from that in the CONV monoculture. Results of this
study suggest that management aimed at increasing cropping system diversity may have
additional effects on weed communities and crop yields beyond those commonly
reported, and these may have important implications for the development of more

efﬁcient and sustainable weed and crop management practices.

46

INTRODUCTION

The relationship between taxonomic diversity and ecosystem functions (e.g.,
productivity, stability) has long been a topic of interest in ecology (MacArthur 1955,
Elton 1958, Cottingham et al. 2001) and may have important implications for the
management of agricultural systems (Altieri 1999). Ecological theory and experimental
evidence from a variety of ecosystems suggest that communities made up of many
interacting species should be less temporally variable (i.e. more stable) than simple
communities made up of few interacting species. This theory also predicts that while
variability at the community level should decrease with diversity, variability at the
population level should increase (Tilman 1996, Lehman and Tilman 2000, Cottingham
et al. 2001) — that is, abundances of individual species may vary over time, but overall
community biomass should remain relatively stable. It is thought that population-level
variability helps maintain aggregate community stability because differential responses
among species to environmental ﬂuctuations compensate for the loss of other species,
thus stabilizing aggregate ecological properties, such as community productivity across
time (Tilman 1996, Micheli et al. 1999, Cottingham et a1. 2001, Tracy and Sanderson
2004). As more growers adopt agronomic practices that increase cropping system
diversity and reduce their reliance on chemical inputs, determining the degree to which
these practices affect the temporal variability of weeds and crop yields may have
important implications for the development of more efﬁcient and sustainable weed and
crop management practices (Helmers et al. 2001, Menalled et al. 2001, Davis et al.

2005)

47

Predicting weed abundance and its effects on crop yield is an important
component of integrated weed management (Lutman et al. 1996). Weed communities
that are less variable may require fewer proactive control measures, such as the use of
broadcast herbicide applications (Maxwell and Luschei 2005). However weed
community variability may also impact other ecosystem processes, such as resource use
efﬁciency (Tracy and Sanderson 2004) and maintenance of natural enemies (Showler
and Greenberg 2003) and these too may impact weeds and crop yields (Altieri 1999).
For example, Thorbek and Bilde (2004) found that highly variable weed communities
did not provide reliable habitat for the maintenance of generalist natural enemy
populations.

While the degree of inter-annual variability in weed communities may have
important implications for the management of weeds and crop yields in agricultural
ecosystems, much of the ecological theory concerning diversity and ecosystem function
has been developed within the context of spatial diversity and in less heavily managed
systems (Schlapfer and Schmid 1999, Cottingham et al. 2001). In many agricultural
systems, however, diversity is both spatial and temporal and little is known about how
the temporal component of diversity (number of different crops or cover crops included
in a rotation) or different agricultural management systems, have on the variability of
weed communities or crop yields.

Here we report results of a study comparing how corn grown in continuous
monoculture and in a 4-yr com-com-soybean-wheat rotation under two input
management systems (conventional and organic-based with cover crops) affects inter-

annual variability in weed communities and crop yields. We tested the hypothesis that

48

aggregate properties of the weed community (total biomass, community structure), as
well as crop yields, should be less variable in more diverse cropping systems that have a
greater number of potential ecological interactions (organic input management systems
and crop rotations) compared to simpler cropping systems with fewer potential
ecological interactions (conventional input systems and continuous monocultures). In
accordance with ecological theory, we also expected that inter-annual variability at the
weed population-level (species richness, species abundance) would be higher in more
diverse cropping systems (Tilman 1996). Variability was assessed at both the
community and population-level using the coefﬁcient of variation (weed community
and population-level and crop yield) and a multivariate dissimilarity index (weed
community-level only). The relationship between inter-annual variability in weed and
crop yield was also assessed.
MATERIALS AND METHODS
Study Site

The study was conducted at the Living Field Lab experimental plots (hereafter
LFL) at the W. K. Kellogg Biological Station Long Term Ecological Research (LTER)
site of Michigan State University in SW Michigan, USA. The LFL was established in
1993. Soils at the LFL site are a mixture of Kalamazoo (ﬁne-loamy, mixed, mesic
Typic Hapludalfs) and Oshtemo (coarse-loamy, mixed, mesic Typic Hapludalfs) sandy
loams (Sanchez et al. 2001). In 1992 the site was planted to alfalfa. In 1993 rotation
treatments were randomly assigned to plots within blocks that were managed with
different input systems. The rotation treatments were continuous corn and each phase

of a four-year com-com-soybean-winter wheat rotation. Only continuous corn and the

49

ﬁrst-year corn phase of the rotation are compared in this study. The input management
systems examined in this study included a conventional system (hereafter CONV) that
received full inputs of fertilizer and herbicides (broadcast applications), at rates
recommended for the region (Sanchez et al., 2004) and an organic-based input system
(hereafter ORG) which received composted dairy manure, but no synthetic chemical
inputs. Each year in the ORG system, crimson clover (T rifolium incarnatum L.) was
sown into standing corn in the continuous corn and the ﬁrst corn phase of the rotation in
late June; Italian ryegrass (Lolium multiﬂorum Lam.) was sown into the standing corn in
the second corn phase; and red clover (T rifolium pretense L.) was frost seeded into
standing winter wheat in late March. No cover crops were sown in the soybean phase
of the rotation.

Weed management in the ORG input system was via periodic (2-4 times/ yr)
inter-row cultivation (row cultivation and rotary hoeing). Plots in both input systems
were chisel plowed and soil ﬁnished each year prior to planting. All treatments (input x
rotation) were replicated four times. Blocks were separated by 8 m grass buffers.
Subplots (rotation treatments) measured 15 x 9 m (CONV) and 15 x 4.5 m (ORG) and
were not separated by buffers. Additional descriptions of the site and management can
be found at: http://1ter.kbs.msu.edu /Data/LTER_Metadata.jsp/Dataset/KBSO42.

Sampling

Each year from 2001 to 2004 weeds were harvested at peak biomass (early
September) from a 0.25 x 1.0 m quadrat placed in each CONV and ORG replicate plot.
Because of the difference in size between the CONV and ORG subplots, two locations

in each CONV and one location in each ORG replicate plot were sampled each year.

50

Within each plot, the quadrat was placed perpendicular to the crop row to ensure equal
sampling of row and inter-row areas. The position of the quadrat within the plot was
varied each year to avoid harvesting in the same location in successive seasons. All
weeds rooted within the quadrat were clipped at ground level, sorted to species, and
dried to constant biomass at 60° C. Dried biomass of each species was weighed to the
nearest 0.01 g. Corn was harvested in October or November each year and yields were
calculated assuming moisture content of 15.5%.
Calculation of Variability

The coefﬁcient of variation (CV) was used to assess inter-annual variability in
weed community and crop yield parameters (Tracy and Sanderson 2004). To assess
aggregate community stability total weed biomass was calculated for each treatment
replicate over the sampling period from 2001-2004. The CV was calculated as the
standard deviation of weed biomass (2001-2004) / average weed biomass (2001-2004).
The CV for quadrat-level species richness (number of weed species occurring in each
quadrat) and corn yield was determined in the same manner. We assessed variability
in species abundances (population-level variability) by calculating the CV for each
species in each treatment replicate. The species CV’s were then averaged for each
replicate (by summing and dividing by the overall number of species present in that
replicate).

In addition to variation in total weed biomass and species richness, we were also
interested in a measure of community variability that would reﬂect overall community
structure (the composition and abundance of species in the weed community). To

obtain a multivariate measure of community variability we calculated a matrix of Bray-

51

Curtis dissimilarity coefﬁcients (McCune and Grace 2002) using a data set containing
species biomass in each treatment replicate for each of the four years. Prior to
calculating the dissimilarity matrix data were log (x + 0.01) transformed and species
occuning in less than 5% of the plot-years were deleted to reduce the inﬂuence of
outliers and rare species. For each treatment replicate we calculated dissimilarity in
community abundance and composition from one year to the next from 2001-2004. The
average annual dissimilarity over the four-year period for each replicate was used as our
measure of community-level variability.
Statistical Analysis

Analysis of variance was used to examine treatment effects on weed community
and crop yield parameters. The model was a split-plot, randomized complete block
with three factors (block, system, and rotation) and the system by rotation interaction.
The block factor was considered random. Tests of main effects were performed with
error terms appropriate for a RCB, split-plot design with input system as the whole plot
factor and rotation as the subplot factor. Treatment means were compared with Tukey’s
HSD test at the 5% level of conﬁdence when main effects were signiﬁcant. All
analyses were performed with SAS (SAS Version 8.02; SAS Institute, Cary, NC, USA)
using the MIXED procedure (Little et al. 1996). Treatment effects on weed biomass
and species richness were also analyzed, and these data were log and square root
transformed, respectively, prior to analysis to improve homoscedasticity.

RESULTS AND DISCUSSION
The effects of the input management system and crop rotation on overall weed

community biomass, richness, and community composition are presented and discussed

52

ﬁrst, followed by presentation and discussion of the treatment effects on weed
community and population-level variability. The third section deals with treatment
effects on crop yields and crop yield variability and their relationship to weed
community variability. Potential explanations for the observed results and their
implications for weed management are discussed in the ﬁnal section.
Weed community biomass and richness

There was substantial variation in precipitation and temperature over the course
of the study (Figure 3.1) and this no doubt contributed to signiﬁcant yearly variation in
the weed community. Weed community biomass and richness varied from year to year
in some treatments by as much as 500 and 300%, respectively (Table 3.1). Averaged
across years, total community biomass was over ten times higher in the ORG compared
to the CONV input system (ANOVA, system: F 1,3 = 186.10, P < 0.001) (Figure 3.2a).
Total weed biomass was unaffected by crop rotation or the interaction between inputs
and rotation (ANOVA, rotation: F 1,6 = 0.04, P = 0.84; interaction: F”, = 0.42, P = 0.54).

There were only marginally signiﬁcant effects of input system and crop rotation
on weed species richness (ANOVA, system: F 1,3 = 8.59, P = 0.06; rotation: F”, = 4.59
P = 0.076; interaction: F 1,6 = 1.35, P = 0.2894 ) (Figure 3.3a). However, the identity of
dominant weed species differed among the treatments (Table 3.2). Abundant weed
species (species making up more than 10% of total weed biomass over the four years) in
the CONV input system included Digitaria ischaemum (Schreb.) Schreb. ex Muhl.
(DIGIS), Solanum carolinense L.(SOLCA), Stellaria media (L.) Vill.(STEME), and
T araxacum oﬂicinale G.H. Weber ex Wiggers (TAROF). In contrast, two annual

dicots, Ambrosia artemisitfolia L. (AMBEL) and Chenopodium album L. (CHEAL),

53

ej— — 2001 30
— - 2002

’e‘ 8’ ' --- 2003 g
s 7r """""""" 2004 E
V ‘ ' a
.5 6* 15 20
‘5 Q)
s s- g
2 8
a 3" c 10
° 8
§ 2 2
2

 

 

1 1 1 1 1 1 1 1
Apr May Jun Jul Aug Sep Oct

1

 

 

l l l L 1
Apr May Jun Jul Aug Sep Oct

1 l 1

Figure 3.1: Mean daily precipitation and air temperature at the KBS LTER for each
month from April-October over the four-year study.

54

Table 3.1: Weed biomass, species richness, and crop yields (mean :t SE; n = 4) in corn
(2001-2004) in conventional (CONV) and organic (ORG) input systems in the KBS
LFL. Rotations are continuous (CC) and rotated (R) com (see text).

 

 

 

 

 

Inputs Rota Year Weed Species Corn
tion biomass richness yield

(g mi) SE (0.25 m'z) SE (bu A") SE

CONV CC 2001 49.0 16.3 5.0 0.8 54.9 4.9

2002 3.4 2.4 3.9 0.4 95.1 1.3

2003 16.5 11.2 4.5 0.8 77.2 6.5

2004 4.9 2.7 5.0 0.6 138.4 5.3

2001 41.7 8.6 7.9 0.8 87.4 6.1

2002 2.6 0.8 7.3 1.8 122.2 2.1

2003 12.2 3.8 4.8 0.4 112.0 2.3

2004 23.5 14.2 4.1 0.9 159.7 7.0

2001 126.5 70.6 6.3 0.6 65.1 4.7

2002 168.1 90.5 5.5 1.0 80.1 13.4

2003 180.7 112.6 5.8 0.5 73.0 4.8

2004 347.8 76.6 10.3 0.8 48.6 6.4

2001 32.1 5.4 6.3 0.3 77.9 10.4

2002 178.5 46.3 10.8 1.5 96.6 10.7

2003 1.0 0.5 3.3 0.5 112.5 3.7

2004 554.2 185.0 9.5 0.7 65.7 11.6

 

55

 

 

 

 

 

400 —
I conﬁnuous

sf '1 rotation
E
U)
(D
(0
E
.9.
f 200
S
o
t—

100

0‘ 1
CONV, ORG
2.0 '-

A conﬂnuous
" rotation

CV

1.0”

 

 

 

0. 5 1 1 1
CONV ORG

Figure 3.2: Effects of conventional (CONV) and organic (ORG) inputs and crop
rotation on a) weed community biomass from 2001-2004 and b) inter-annual variability
as measured by the coefﬁcient of variation, CV. Error bars are i 1 SE; total biomass, n
= 16; CV of biomass, n = 4. Bars sharing the same letter are not signiﬁcantly different
at the P = 0.05 level (Tukey-Kramer Test).

56

(a)

I continuous
3 _ 1‘ rotation

Species richness
O)
l

 

 

 

 

 

CONV ORG

(b)

0.7 — .
A conﬁnuous

rotation
0.6 —

0.5 “

CV

0.4 '—

 

 

02- i

01 ' ' '
CONV ORG

 

 

Figure 3.3: Effects of conventional (CONV) and organic (ORG) inputs and crop
rotation on a) weed community species richness from 2001-2004 and b) inter-annual
variability as measured by the coefﬁcient of variation. Error bars are i 1 SE; richness,
n = 16; CV of richness, n = 4. Bars sharing the same letter are not signiﬁcantly
different at the P = 0.05 level (Tukey-Kramer Test).

57

Table 3.2: Weed species abundance in conventional (CONV) and organic (ORG) input
systems under continuous (CC) and rotation (R) production. Values are mean % of total
weed biomass and CV averaged across years (2001-2004).

 

 

 

 

CONV ORG

Biomass CV Biomass CV
Species (%) CC R (%) CC R
Abutilon theophrasti 0.08 0.50 0.43
Amaranthus retroﬂexus 0.04 0.45 7.97 0.48 0.50
Ambrosia artemisiifolia 30.40 0.43
Arabidopsis thaliana 0.60 0.35 0.40 0.05 0.43 0.40
Cardamine sp. 0.01 0.50
Chen0podium album 0.65 0.30 0.26 35.06 0.37 0.46
Conyza canadensis 0.00 0.50
Daucus carota 0.05 0.50
Digitaria ischaemum 14.98 0.50 0.1 7 0.46
Digitaria sanguinalis 1.20 0.37 0.34 0.03 0.45 0.43
Echinochloa crus-galli 0.81 0.37 0.50
Elymus repens 2.29 0.50 0.50 6.55 .40 0.47
Eragrostis cilianensis 0.00 0.50 0.00 0.50
Erigeron sp. 0.16 0.50
Lamium sp. 0.08 0.36 0.50 0.03 0.36 0.43
Mollugo verticillata 0.01 0.50
Oxalis stricta 0.08 0.46 0.04 0.50
Panicum dichotomiﬂorum 0.87 0.45 0.00 0.50
Plantago major 0.03 0.50 0.25 0.50
Poa sp. 0.68 0.26 0.30 0.01 0.33 0.50
Polygonum aviculare 0.09 0.50 0.50
Polygonum convolvulus 1.12 0.36 0.50
Polygonum persicaria 2.35 0.44 0.50
Portulaca oleracea 0.03 0.50 0.40 0.00 0.50
Rumex obtusifolius 1.58 0.50 0.38 0.01 0.50
Setariafaberi 0.02 0.50 2.68 0.50
Setaria pumila 0.01 0.50 1.61 0.50 0.50
Setaria viridis 0.08 0.50 3.15 0.50 0.46
Solanum carolinense 17.39 0.47 0.41 1.15 0.49 0.50
Solanum ptychanthum 0.00 0.50 0.05 0.50 0.48
Sonchus oleraceus 0.07 0.50
Stellaria media 19.13 0.43 0.45 2.31 0.44 0.48
T araxacum oﬂicinale 37.83 0.38 0.39 1.11 0.24 0.42
Veronica persica 0.05 0.50 0.06 0.50 0.40
Veronica peregrina 0.41 0.49 0.50 0.03 0.50 0.47

 

58

were the most abundant in the ORG input system, making up 30 and 35%, respectively,
of the total biomass.

We had expected that these input systems would vary in total weed biomass
because mechanical control measures are often less effective than chemicals at
controlling weeds, particularly later in the growing season when crop canopy
architecture precludes mechanical cultivation (Pleasant et a1. 1994). Given the number
of studies that have reported signiﬁcant rotation effects on weed abundance and
diversity (Liebman and Dyck 1993), the absence of a rotation effect in our study is
somewhat surprising; though not unique. Doucet et al. (1999) found few residual
effects after nine years of a com-soybean-wheat rotation on the composition and
abundance of weeds following planting of corn in the 10th year. Similarly, in an earlier
study in the LF L, the previous crop, rather than the number of different crops in the
rotation was found to have the greatest impact on the abundance and composition of
seeds in the soil (Smith and Gross, in review).

Weed community variability

We used a number of different metrics to characterize the level of inter-annual
variability in the weed community in response to the input management system and
rotation treatments and although we detected differences among the treatments, they
were not consistent with the predictions we had developed based on diversity-stability
theory (Schléipfer and Schmid 1999, Cottingham et al. 2001). For example, based on
ecological theory we predicted that the more diverse rotation (ORG rotation) would be
less prone to ﬂuctuations in weed biomass. However, we found that community-level

variability was higher in the ORG rotation compared to the ORG monoculture

59

(ANOVA, interaction: F H, = 26.09, P = 0.002, Tukey-Kramer, P = 0.0188) (Figure
3.2b). Similar results were observed when we analyzed weed community variability
using a dissimilarity metric that combined data on both species abundance and
composition (Bray-Curtis distance). The analysis indicated that weed communities
were signiﬁcantly more dissimilar in abundance and composition from year to year in
the ORG rotation than the ORG or CONV monoculture or CONV rotation (system: F. ,3
= 20.97, P = 0.02; rotation: Fm, = 13.28, P = 0.011; interaction: F”, = 5.56, P = 0.056)
(Figure 3.4a). Taken together these results suggest that both input management and
crop rotation can affect weed community variability but that crop rotation may be a
more important driver of variability. Our results are rather similar to those of Davis et
al. (2005) who reported little difference in the constancy of weed communities in
conventional and organic rotations. This is in contrast, however, to a similar study
conducted by Menalled et al. (2001) that reported weed communities in organic systems
were more diverse and more constant compared to conventional-input systems,
suggesting that it may be difﬁcult to make general predictions about how input
management will affect temporal variability in weed communities (Davis et al. 2005).
Diversity-stability theory predicts that as the number of species in a community
increases, variability in the abundance of individual species should also increase
(Tilman 1996). We found this to be true; weed p0pulation-level variability (average CV
for individual species) was highest in the most diverse cropping system (ORG rotation)
and lowest in the least diverse cropping system (CONV monoculture) (system: F13 =
10.18, P < 0.05; rotation: F”, = 8.01, P = 0.03, interaction: F”, = 0.13 P = 0.73) (Table

3.2 and Figure 3.4b). However, weed population and community level variability did

60

(a)

 

 

 

 

 

 

 

 

1.0 ~
I continuous b
1 rotation
0.9 r +
g
g 0.8 ’— a
.Li a
.8
D 0.7 - a
0.6 -
05 '
CONV ORG
0'50 F A continuous
’1?- rotation
0.45 -
> I
O
0.40 r
0.35 I ' ‘

 

 

CONV ORG

Figure 3.4: Effects of conventional (CONV) and organic (ORG) inputs and crop
rotation on a) weed community dissimilarity and b) population-level inter-annual
variability as measured by the coefﬁcient of variation. Error bars are i 1 SE; n = 4.

Bars sharing the same letter are not signiﬁcantly different at the P = 0.05 level (Tukey-
Kramer Test).

61

not appear linked in the manner predicted by diversity-stability theory, as community
level-variability was also higher in the ORG rotation (Figs. 3.2b and 3.4a). This
suggests that the link between weed population and community-level stability may be
weak in systems that are highly disturbed, such as agricultural ﬁelds that are annually
tilled (Chesson and Huntly 1997).

Variability in community richness (CV richness) was not affected by either of
the treatments (only a marginally signiﬁcant effect of rotation: F .1, = 4.91, P = 0.07)
(Figure 3.3b).

Corn yields and variability

Corn yields also varied among years (Table 3.1). Averaged across the four
years, corn yields were highest in the CONV rotation, intermediate in the CONV
monoculture and ORG rotation, and lowest in the ORG monoculture (system: F. ,3 =
19.81, P = 0.02; rotation: F”, = 168.5 P < 0.0001, interaction: F”, = 3.08, P = 0.13)
(Figure 3.5a). As predicted from diversity-stability theory, we found that inter-annual
variability in corn yield was highest (highest CV) in the simpler cropping system
(CONV) (system: F .,3 = 12.44, P = 0.04; rotation: F .1, = 0.54, P = 0.49, interaction: F. ,6
= 2.10, P = 0.20) (Figure 3.5b). This result is in contrast to that of Tracy and Sanderson
(2004) who found no relationship between the diversity of forage species and inter-
annual yield variability.

We found no relationship between variability in corn yields and variability in the
weed community using any of the weed community variability indices as independent

variables in a linear regression (data not shown).

62

FIT I continuous
’ ' rotation

 

Yield (Mg ha")

 

 

 

 

 

 

CONV ORG

(b)

0.6
F A continuous

rotation

0.5 — §

0.4 "

CV

0.3 *

 

 

 

0.1 1 1 1
CONV ORG

Figure 3.5: Effects of conventional (CONV) and organic (ORG) inputs and crop
rotation on com a) yields from 2001-2004 and b) inter-annual variability as measured
by the coefﬁcient of variation. Error bars are i 1 SE; yield, 11 = 16; CV, n = 4. Bars
sharing the same letter are not signiﬁcantly different at the P = 0.05 level (Tukey-
Kramer Test).

63

Potential explanations and implications

One potential explanation for the lower inter-annual variability in weed
communities in the continuous corn compared to the rotation could be the fact that
within a block, continuous corn is planted in the same location each year. This is in
contrast to rotated corn, which, because each phase of the rotation is present each year,
occurs in a different location within the block each of the four years. If spatial
differences in weed species abundance between continuous corn and rotated corn were
responsible for the observed differences in inter-annual variation then one would expect
that within years, differences among replicates would be greater in the rotated corn
treatments compared to continuous corn. To test this, we compared mean intra-year
dissimilarity among replicate plots for each treatment using Bray-Curtis dissimilarity
coefﬁcients. We found that annual dissimilarity between replicates did not differ
among the treatments (ANOVA, system: F.,3 = 0.12, P = 0.75; rotation: Fm = 3.21 P =
0.123; interaction: F. ,6 = 0.36, P = 0.571 ), indicating that temporal autocorrelation was
likely not responsible for the observed effects of rotation on community variability.

Another potential explanation may be that effects of diversity at the scale of the
crop are independent of those at the scale of the weed community. This may be
especially true when crop diversity is temporal (as in our crop rotations) and punctuated
by high rates of annual disturbance. High rates of annual disturbance can lead to weak
interactions among species (Chesson and Huntly 1997), and this may affect the
compensatory responses among weed species necessary for generating stability

(Cottingham et al. 2001).

64

If cropping system diversity per se has little direct effect on weed community
variability, how can the differences in variability among the rotation treatments be
explained? One possible answer may be the soil seed bank. Seeds in the seed bank
contribute to most of the weeds emerging in a ﬁeld in any given year and the age of
seeds in the seed bank can vary (Cavers and Benoit 1989). Seed age has been shown to
affect germinability (Lindgren and Schaaf 2004), and can affect recruitment in some
species (Rice and Dyer 2001). In our study, many of the weeds that emerged each year
in the corn phase of the rotation likely germinated from seeds that were shed the last
time corn appeared in the rotation (i.e. > 2 years earlier). This is in contrast to
continuous com, in which the majority of weeds likely emerged from seeds shed the
previous year (Cavers and Benoit 1989). The seed bank in rotated corn also likely
contained an abundance of species associated with winter wheat (Smith and Gross, in
review), which might also have contributed to compositional variation.

Our results suggest that in addition to the environmental costs often associated
with conventional agricultural management (Robertson and Swinton 2005) crop yields
in conventionally managed systems may be more prone to inter-annual variability
(Helmers et al. 2001). This is in contrast to the organic rotation, in which yields were
signiﬁcantly less variable, and were just as high as those in the conventional
monoculture, despite this system having greater weed abundance and weed community
variability. This suggests that either the interactions among crops and weeds are lower
in diversiﬁed organic input systems (Stevenson et al. 1998) or that some other property

of the weed community, which we did not measure, contributes to increased crop yield

65

stability. A better understanding of the factors contributing to temporal yield variation

in row-crops will help answer this question.

ACKNOWLEDGMENTS

We acknowledge and thank Dr. R. R. Harwood for allowing us to sample the
Living Field Lab (LFL) and for having the foresight to design and establish this
experiment. We also thank G. Parker (farm manager) and J. Smeenk for their work in
maintaining these treatments. C. Baker, M. Hammond, and numerous lab assistants
helped with ﬁeld sampling. Drs. D. D. Buhler, D. A. Landis, C. A. Malmstrom, and G.
G. Mittelbach provided helpful comments on earlier drafts of this chapter. Support for
this study was provided by the NSF-funded KBS LTER project (DEB98-10220; KLG)

and by a fellowship from the CS. Mott Program in Sustainable Agriculture (RGS).

66

LITERATURE CITED

Altieri, M. A. 1999. The ecological role of biodiversity in agroecosystems. Agriculture
Ecosystems and Environment 74: 19-31.

Cavers, P. B., Benoit, D. L. 1989. Seed banks in arable land. In: Ecology of Soil Seed
Banks, (eds. MA Leck, VT Parker & RL Simpson), 309-328. Academic Press,
San Diego, USA.

Chesson, P., Huntly, N. 1997. The roles of harsh and ﬂuctuating conditions in the
dynamics of ecological communities. American Naturalist 150:519-553.

Cottingham, K. L., Brown, B. L., Lennon, J. T. 2001. Biodiversity may regulate the
temporal variability of ecological systems. Ecology Letters 4:72-85.

Davis, A. S., Renner, K. A., Gross, K. L. 2005. Weed seedbank and community shifts in
a long-term cropping systems experiment. Weed Science 53: 296-306.

Doucet, C., Weaver, S. E., Hamill, A. S., Zhang, J. H. 1999. Separating the effects of
crop rotation from weed management on weed density and diversity. Weed
Science 47:729-735.

Elton, C. 1958. The ecology of invasions by animals and plants. Chapman & Hall,
London, UK.

Helmets, G. A., Yamoah, C. F., Varvel, G. E. 2001. Separating the impacts of crop
diversiﬁcation and rotations on risk. Agronomy Journal 93:1337-1340.

Lehman, C. L., Tilman, D. 2000. Biodiversity, stability, and productivity in competitive
communities. American Naturalist 156:534-552.

Liebman, M., Dyck, E. Crop rotation and intercropping strategies for weed
management. Ecological Applications 3: 92-122.

Lindgren, D.T., Schaaf, D. M. 2004. Inﬂuence of seed stratiﬁcation and seed age on
emergence of Penstemon. HortScience 39:1385-1386.

Little, R. C., Milliken, G. A., Stroup, W. W., Wolﬁnger, R. D. 1996. SAS system for
mixed models. SAS Institute, Cary, North Carolina, USA.

Lutman, P. J. W., Risiott, R., Ostennann, H. P. 1996. Investigations into alternative
methods to predict the competitive effects of weeds on crop yields. Weed
Science 44: 290-297.

MacArthur, R. 1955. Fluctuations of animal populations, and a measure of community
stability. Ecology 36:533-536.

67

Maxwell, B. D., Luschei, E. C. 2005. Justiﬁcation for site-speciﬁc weed management
based on ecology and economics. Weed Science 53: 221-227.

Menalled, F. D., Gross, K. L., Hammond, M. 2001. Weed aboveground and seedbank
community responses to agricultural management systems. Ecological
Applications 1 121586-1601.

Micheli, F., Cottingham, K. L., Bascompte, J ., Bjomstad, O. N., Eckert, G. L., Fischer,
J. M., Keitt, T. H., Kendall, B. E., Klug, J. L., Rusak, J. A. The dual nature of
community variability. Oikos 85:161-169.

Rice, K. J ., Dyer, A. R. 2001. Seed aging, delayed germination and reduced competitive
ability in Bromus tectorum. Plant Ecology 155: 237-243.

Robertson, G. P., Swinton, S. M. 2005. Reconciling agricultural productivity and
environmental integrity: a grand challenge for agriculture. Frontiers in Ecology
and the Environment 3: 38-46.

Sanchez, J. E., Willson, T. C., Kizilkaya, K., Parker, E., Harwood, R. R. 2001.
Enhancing the mineralizable nitrogen pool through substrate diversity in long
term cropping systems. Soil Science Society of America Journal 65:1442-1447.

Schlapfer, F., Schmid, B. 1999. Ecosystem effects of biodiversity: a classiﬁcation of
hypotheses and exploration of empirical results. Ecological Applications 9: 893-
912.

Showler, A. T., Greenberg, S. M. 2003. Effects of weeds on selected arthropod
herbivore and natural enemy populations, and on cotton grth and yield.
Environmental Entomology 32:39-50.

Stevenson, F. C., Legere, A., Simard, R. R., Angers, D. A., Pageau, D., Lafond, J. 1998.
Manure, tillage, and crop rotation: Effects on residual weed interference in
spring barley cropping systems. Agronomy Journal 90: 496-504.

Tilman, D. 1996. Biodiversity: Population versus ecosystem stability. Ecology 77:350-
363.

Thorbek, P., Bilde, T. 2004. Reduced numbers of generalist arthropod predators after
crop management. Journal of Applied Ecology 41 :526-538.

Tracy, B. F ., Sanderson, M. A. 2004. Productivity and stability relationships in mowed

pasture communities of varying species composition. Crop Science 44:2180-
2186.

68

CHAPTER 4
WEED COMMUNITY RESPONSE TO INCREASING CROP DIVERSITY
ABSTRACT
Increasing diversity in cropping systems may have important implications for
agroecosystem functions related to the regulation of weed populations. However, the
impacts of cropping system diversity per se on weed communities have not been
formally evaluated in a study comparable to those performed in grassland systems,
where much of the evidence regarding diversity-ecosystem function has been reported.
Here we report results from three yrs of a long-term study designed to examine the
effects of crop diversity on the abundance, diversity, and composition of the weed
community. Crop diversity was manipulated through rotations of corn (Zea mays L.),
soybean (Glycine max (L.) Merr.), winter wheat (T riticum aestivum) and cover crops.
Weed community abundance and composition was measured at peak biomass (August-
September) for three years from 2002-2004; soil resources and light levels were
measured in 2004. Weed abundance and diversity were both affected by crop diversity;
however, effects were stronger in winter wheat than corn and soybean. In winter wheat
weed abundance and diversity (species richness, H’ and D) were lowest in the two
highest diversity treatments. The effects of crop diversity in wheat were due mainly to
the presence of cover crops in the higher diversity treatments and likely reﬂected the
impacts of cover crops on soil resource and light levels. The effects on weed
community structure of crop diversity per se were slight compared to the effects of crop
identity. Crop rotation had little persistent effect on weed abundance and composition

despite large differences among crops. This study suggests that increasing crop

69

diversity may decrease the need for weed management, not because of its effects on
weed abundance and composition, but because of its effects on resource availability and
crop competitiveness.

INTRODUCTION

Increasing cropping-system diversity may have important implications for the
functioning of agricultural ecosystems (Altieri 1999, Minns et al. 2001) and has been
advocated as a potential means of decreasing the need for intensive chemical inputs for
weed control (Dekker 1997, Kegode et al. 1999, Westerman et al. 2005, Liebman and
Gallandt 1997, Liebman and Staver 2001). However, ecological context can affect
diversity-ecosystem function relationships (Schlapfer and Schmid 1999) and much of
the experimental evidence regarding diversity and ecosystem function has been reported
from studies conducted in relatively undisturbed natural and synthetic grassland systems
where biotic conditions and disturbance regimes vary from those in most managed
systems (Lambers et al. 2004, Hooper et al. 2005, Loreau et a1. 2001, Tilman et al.
2001, Levine and D’ Antoino 1999, Cottingham et al. 2001). It is therefore unclear
whether and to what extent the relationships between community diversity and
ecosystem function commonly reported in experimental grasslands can be applied to
managed systems, such as row-crop agroecosystems.

Studies conducted in grasslands differ from those in annual row-crop
agroecosystems in several ways that might impact the generality of diversity-ecosystem
function relationships and their application to weed management. In synthetic grassland
communities, manipulations of diversity include a relatively large number of species

and potential species combinations (Schlapfer and Schmid 1999; Hooper et al. 2005).

70

Diversity manipulations in these systems also tend to span relatively large diversity
gradients (Tilman et al. 1997, Hooper and Vitousek 1998, Hector et a1. 1999), with
treatment levels in some experiments ranging from one to as many as 32 species
(Hector et al. 1999). In addition to having many levels of diversity, the densities of
individuals in grassland manipulations tend to be high. High densities coupled with low
disturbance rates following initial community establishment allow the potential for
strong interactions among species (Cousens 1985). Inter and intra-speciﬁc interactions
are hypothesized to drive many of the observed relationships between diversity and
ecosystem function and underlie many of the mechanistic predictions regarding
diversity’s effect on community productivity and the establishment of new species
(Tilman et al. 1997, Schlapfer and Schmid 1999, Lambers et al. 2004, Hooper et al.
2005)

In contrast to grassland systems, studies of diversity in most agricultural systems
typically include a limited number of crop species and potential species combinations.
Because of the low numbers of crops and realistic cropping combinations, the gradients
of diversity manipulated in many agricultural studies tend to be narrow and often only
include a few levels of crop diversity (Jolliffe 1997, Schlapfer and Schmid 1999).

Row spacing and planting densities are controlled in most agricultural systems,
resulting in relatively low densities of individuals compared to grassland systems. Most
agricultural systems are also highly disturbed by management practices such as
cultivation, and the applications of fertilizer and herbicides; and these can also affect the
density of individuals and the strength of species interactions (Huston 1979, Wiener et

al. 2001, Cardinale and Palmer 2002). In addition, a large component of diversity in

71

 

most agricultural systems is temporal (crop rotation) (Liebman and Dyck 1993,
Liebman and Staver 2001), meaning that many of the interactions among crop species
are indirect, in contrast to grassland systems where many species interact directly
(Tilman 1999).

There have been many studies comparing weed communities in relatively simple
and complex cropping systems (Liebman and Dyck 1993, Barberi and Lo Cascio 2001,
Legere and Stevenson 2002, Cardina et al. 2002, Moonen and Barberi 2004, Teasdale et
al. 2004, Westerman et al. 2005). However, there have been no large-scale
manipulations of diversity in agricultural systems that are comparable to those
performed in grassland systems. Additionally, there have been no studies of diversity in
agricultural systems that span a relatively broad range of crop diversity and that do not
include management practices such as fertilizer inputs and herbicide applications
(Liebman and Dyck 1993, Schlapfer and Schmid 1999), which can potentially confound
or obscure the effects of diversity per se (Doucet et a1. 1999).

Here we report data from the ﬁrst three years (2002-2004) of a long-tenn
cropping-system diversity study established in 2000 to examine the effects of crop
diversity on ecosystem functions and processes related to row crop agriculture. For this
paper we focus on how crop diversity inﬂuences the abundance, diversity, and
composition of the emergent weed community over a three-year period from 2002 to
2004. Crop diversity was manipulated through rotations of corn, soybean, winter wheat
(main crops) and cover crops. After 2001 the experiment was conducted in the absence
of external chemical inputs (chemical fertilizer, herbicides, and pesticides), so that

diversity effects could be separated from those of speciﬁc management practices that

72

can vary with each crop. In this study we examined ﬁve main questions: (1) Does
increasing crop diversity affect the overall abundance of weeds?, (2) Does crop
diversity affect the diversity, composition, and structure of weed communities ?, (3) Do
the effects of crop diversity on the weed community differ depending on the identity of
the main crop?, (4) What is the relative importance of cover crops versus crop rotation
on crop diversity-ecosystem function relationships, and (5) Are the effects of crop
diversity on weed communities related to changes in soil resource levels and/or the light
environment?
MATERIALS AND METHODS
Study site

The study was conducted at the Biodiversity Experiment Plots (hereafter BEP)
at the W. K. Kellogg Biological Station (KBS) Long Term Ecological Research (LTER)
project in agricultural ecology in Hickory Comers, Michigan, USA. Soils at the study
site are dominated by Kalamazoo silt loam (Typic Hapludalfs), and are made up of 43%
sand, 40 % silt, and 17 % clay (Robertson et al. 1997). Mean annual temperature at the
LTER site is 9.7°C. Annual precipitation is 890 mm], with about half in the form of
snow (http://lter.kbs.msu.edu/siteDescription.html).

Diversity treatments

The BEP was established in 200010 examine the effects of realistic
manipulations of row crop diversity on ecosystem functions related to row crop
agriculture. Treatment systems consist of three row-crops, corn (Zea mays L.), soybean
(Glycine max (L.) Men), and winter wheat (T riticum aestivum L.) grown in continuous

monoculture and in two- and three-year annual rotations with and without cover crops

73

(zero, one, or two legume/small grain species). The six crop diversity treatments range
in diversity from one to three crop species annually and one to six species over a three-
year period (Table 4.1) and include: (1) continuous monoculture (1 spp), (2) continuous
monoculture with one cover crop annually (2 spp), (3) two-crop rotation (2 spp), (4)
three-crop rotation (3 spp), (5) three-crop rotation with one cover crop annually (2 spp
annually; 4 spp over 3 yr rotation), and (6) three-crop rotation with two cover crop
species annually (3 spp annually; 6 spp over 3 yr rotation). In the two-crop rotation
(treatment 3), only three of the possible sequences were included in the design: com-
soybean, soybean-com, and soybean-wheat. The three-crop rotations (treatments 4-6)
are all planted in the same sequence: com-soybean-wheat. Each phase of the two and
three-crop rotations is present every year allowing comparisons to be made among the
six diversity treatments for each of the three main crop species.

Cover crops used in the experiment vary depending on the main crop and
include red clover (T rifolium pratense L.), crimson clover (T rifolium incarnatum L.)
and cereal rye (Secale cereale L.). In treatments receiving a single cover crop
(treatments 2 and 5), T. pretense is sown into corn in July and wheat in March, and S.
cereale is sown in soybean (treatment 2 only) in October. For the highest diversity
treatment (treatment 6) each crop is grown with two cover crops: T. pratense is planted
into corn and wheat as above, and T. incarnatum is planted into soybean in July. Secale
cereale is sown into corn and wheat stubble in October.

Diversity treatments were randomly applied to plots in three linear blocks; in the

fourth block, treatments were arranged in sequence in a 4 x 5 block to facilitate

74

Table 4.1: Crop diversity treatments at the KBS LTER Biodiversity Experiment Plots
(BEP). Treatments were established in 2000 and include monocultures and two- and
three-crop rotations of corn, soybeans and wheat, with and without grass and legume
cover crops.

 

 

 

 

Treatment Description No. crop species
Group #
Annual Sum over rotation
Crops Cover Total Crops Cover Total
crops crops
1 Continuous monoculture l 0 1 1 1
2 Continuous monoculture, 1 1 2 1 1

one cover crop annually

3 Two-crop rotation 0 1 2 0 2
4 Three-crop rotation 0 1 3 0 3
5 Three-crop rotation, one 1 3 2 5

cover crop annually
6 Three-crop rotation, two 1 2 3 3 3 6
cover crops annually

75

demonstration of treatment effects. Replicate blocks were separated by 8m grass
buffers. Individual plots measure 9.1m x 27.4 m with no buffer strips between plots.
All plots are chisel plowed and soil ﬁnished prior to planting the main crop. The exact
timing of planting each year is dependent on weather conditions. Com and soybean are
planted in late April or May, respectively, and winter wheat is planted in late September
or early October. No external inputs (herbicides, pesticides, or fertilizer) were applied
to any of the treatments after 2001; however, periodic inter-row cultivations
(cultivation/rotary hoeing) were performed in corn and soybean each year early in the
growing season.
Vegetation sampling

Aboveground weed and cover crop biomass was harvested at peak biomass
(August-September) from two 0.25 x 1 m quadrats placed in two central locations in
each plot. Weed species composition was determined from a 0.5 x 2 m quadrat that
included the area harvested. The sampling location was changed each year to minimize
edge effects and avoid harvesting weeds from areas that had been harvested the
previous season. Harvested weed biomass was separated to species, dried at 65° C for
48 hrs and weighed to the nearest 0.01 g.

Soil Resources and light

To determine the relationship between crop diversity, weed communities and
soil resource levels, we measured inorganic soil nitrogen (NO3' and NH4+) at the end of
the growing season (November) in 2002 and at three times: 28 June (mid-season), 23
August (late-season), and 16 November (post-season) in 2004. Soils were sampled to a

depth of 25 cm (four-2 cm diameter cores) within the central 3 x 9 m of each plot from

76

both the rows and inter-rows. Aggregated samples were sieved through a 4 mm screen
to remove stones and large pieces of organic material. Following sieving, a portion of
each aggregated sample (20-30 g) was processed for gravimetric soil moisture analysis.
The remaining sample was subdivided; 20 g was processed for N extraction with 100 m1
of 1M KCL (20 g) and the rest air-dried and archived. Nitrogen (NO3’ and NHX) was
analyzed with an Alpkem 3550 continuous-ﬂow colorimetric analyzer (01 Analytical,
College Station, TX, USA). Soils were stored at 4 ° C and were processed within 48 hrs
of collection. In addition, soils sampled in 2002 and on 16 November in 2004 were also
analyzed for net-mineralizable N following incubation for 28 days at 25 ° C.

To determine how the manipulations of crop diversity affected the light
environment throughout the growing season, the percentage of Photosynthetically
Active Radiation (PAR) at the soil surface was measured with an AccuPar LP-80
Ceptometer (Decagon Devices, Inc., Pullman, WA, USA) on three dates in 2004: 23
July (mid-season), 22 August (late-season), and 19 September (post-season). PAR was
measured at solar noon :t 2 hrs by placing the ceptometer parallel to the crop row at two
locations within the center 3 x 5 m of each plot. The two readings were averaged to
represent overall plot-level PAR.

Statistical analyses

Analysis of variance was used to test for the effects of crop diversity, main crop
identity, and year of sampling on weed biomass, species richness (number per m2), and
several measures of community diversity. The initial model included four factors
(block, diversity, crop, and year) and the factorial combinations of crop diversity, crop,

and year. Data were analyzed as a repeated measures, randomized compete block

77

design using PROC MIXED (Little et al. 1996) in SAS (SAS Version 8.02; SAS
Institute, Cary, NC, USA). Separate analyses were performed for each crop and for
each crop x year when higher order interactions were signiﬁcant. Treatment means
were compared with Tukey’s HSD test at the P < 0.05 level of probability when main
effects were signiﬁcant. In all analyses, the block effect was considered random.

We chose to investigate several different diversity measures because each
emphasize different aspects of community structure (Magurran 1988) and the predicted
effects of crop diversity on weed community structure are somewhat unclear (Liebman
and Dyck 1993, Liebman and Staver 2001). The diversity measures were: Shannon
diversity: H’ = -Zp.(log p.), where p. is the proportion of biomass accounted for by
species i per sample (Magurran, 1988), Evenness: E = H’/ln(species richness)(Pielou,
1969, p. 233), and Simpson’s index of diversity: D = 1 — 2p.2 (McCune and Grace
2002). Weed biomass and species richness data were log and square root transformed,
respectively, to improve homoscedasticity. Analysis of variance also was used to
analyze treatment effects on soil resource levels and percent active radiation (PAR).

Several multivariate analyses were employed to examine treatment effects on
the weed community and to visualize differences in community composition and
structure among the six crop diversity treatments. Prior to performing analyses on
species abundance and composition data, all values were log (x + 1) transformed. A
dissimilarity matrix using Bray-Curtis coefﬁcients was then calculated from the
transformed species abundance values. Unknown and very rare species (those

occurring in fewer than 5% of the plot-years) were deleted prior to analysis.

78

The ﬁrst analysis investigated treatment effects on community abundance and
composition in each crop using a distance-based multivariate analysis for a linear model
(DISTLM v.5, Anderson 2001, 2004). The data set contained log-transformed weed
species abundance data collected from 2002 through 2004. For the multivariate linear
model, three factors (block, crop diversity, and year) and the interaction between crop
diversity and year were analyzed using a randomized complete block design with
repeated measures. Year was the repeated factor and block was treated as a random
factor. Treatment effects were tested by permutation (9999 permutations) using error
terms appropriate for a randomized complete block design. A Monte Carlo sampling
procedure was then used to generate P -values from the asymptotic permutation
distribution (Anderson 2004).

Forward multivariate multiple regression (McArdle and Anderson 2001,
Anderson 2003) was performed to characterize species abundance and composition in
2004 in relation to resource levels (inorganic N, soil moisture, and %PAR) measured
during the2004 growing season. Species abundance data were log (x +1) transformed
prior to calculating Bray-Curtis dissimilarity coefﬁcients. For each crop, marginal tests
were performed on all predictor variables ﬁrst, followed by forward selection of
variables with conditional tests. Signiﬁcance was assessed by permutation (9999
permutations) (Anderson 2003).

To help visualize treatment effects on community structure, we performed non-
metric multidimensional scaling ordination (NMDS) (McCune and Grace 2002) on log
(x + 0.1) transformed species composition and abundance data from 2004. Ordination

was performed for each crop separately so that differences among treatments would be

79

more easily detected. Pearson correlation coefﬁcients and r2 between the NMDS
ordination axes scores and soil resource and light levels were also calculated to examine
the strength of the relationship between community structure on soil resources and light
using the PC-ORD software package (McCune and Mefford 1999).

For each main crop, species-level responses to the diversity treatments were
analyzed with indicator species analysis (ISA; Dufrene and Legendre 1997) using
diversity system as the grouping variable. Indicator values (IV) were calculated by
multiplying the relative abundance and relative frequency of each species in each
diversity treatment in 2004. For each species, the signiﬁcance of the highest indicator
value (corresponding to the diversity system with which it was most associated) was
tested with a Monty Carlo procedure (1000 permutations) at a the P < 0.05 level of
probability. ISA was also performed on presence/absence data across diversity
treatments, using crop identity as the grouping variable, to determine the strength of
species associations for each main crop.

RESULTS
Weed biomass

Weed biomass was affected by crop diversity, crop identity, year of sampling,
and the interaction between all three factors (3-way interaction: F29,.09 = 4.36, P <
0.000; Figure 4.1). Subsequent repeated measures analyses for each main crop showed
that most of the effects across crops were driven by a signiﬁcant crop diversity-by-year
interaction in wheat (F933 = 13.15, P < 0.0001 ). In all three years, wheat grown in
treatments that did not include cover crops (i.e. treatments 1, 3, and 4) had higher weed

biomass than wheat grown in treatments that included cover crops (treatments 2, 5, and

80

   

 

 

5.: 2— Treatment
9 01
81 1— 02
:3 413
8 0— '74
s '5
:5 '6
'O ‘1—
(D
o
E

-2—

1 1 1 1

 

 

2002 2003 2004
Year

Figure 4.1: Weed biomass in relation to crop diversity over time in the KBS
Biodiversity Experiment Plots. Symbols refer to crop diversity treatments (see Table
4.1). Symbols that are shaded indicate treatments with cover crops. Values are means
(:1: SE; n = 12) averaged across the three crops.

81

 

 

 

2 m Q) .9
0E 2 — 1;: ' “ :5: Treatment
92 O 1
8) 1 — O 2
:1. ’-> 3
(D \‘7
8 0 __ 4
g v 5
'5' I 6
'O '1 “—
0
é’

-2 _

_3 1 1 L 1

2002 2003 2004

Year

Figure 4.2: Weed biomass in winter wheat planted under different crop diversity
treatments (2002-2004). Symbols refer to crop diversity treatments (see Table 4.1).
Symbols that are shaded indicate treatments with cover crops. Values are means (i SE;
n = 4).

82

 

6; Tukey test P > 0.05; Figure 4.2). While there was a signiﬁcant effect of year on
weed biomass in both corn and soybean (corn: P = 0.02; soybean; P < 0.001), there was
no signiﬁcant crop diversity or crop diversity-by-year interaction for either crop (data
not shown).
Species richness and diversity

Weed species richness was affected by the interaction of crop identity, crop
diversity, and year (3-way interaction: F 29.09 = 4.36, P < 0.0001) and overall was lower
in the lowest diversity treatments (Figure 4.3). Subsequent repeated measures analysis
revealed signiﬁcant crop diversity-by-year interactions in all three main crops (2-way
interaction, corn: F.o,3(, = 2.99, P = 0.008; soybean: Fm...) = 4.96, P = 0.0001; wheat:
F 9,33 = 8.82, P < 0.0001). In corn, differences among treatments were signiﬁcant in
2003 (P = 0.0168) and 2004 (P = 0.0378); however, an adjusted multiple-comparison
test indicated that in 2003 only diversity treatments 2 and 4 (continuous monoculture
with cover crop and the three-crop rotation, respectively) were signiﬁcantly different
from the others; none of the diversity treatments were different in 2004 (Tukey HSD
test, P < 0.05). In soybean, differences among diversity treatments were signiﬁcant
only in 2003; treatments 5 and 6 (three-crop rotation with one and two cover crop
species, respectively) differed from treatment 3 (two-crop rotation), Tukey HSD test, P
< 0.05). In contrast, there were large differences in weed species richness among
treatments planted to wheat in all years, and the differences among diversity treatments
increased over time (Figure 4.4). In general, wheat grown in treatments that included
cover crops (treatments 2, 5, and 6) had lower weed species richness compared to

treatments without cover crops. Species richness did not differ between treatments 1, 3

83

   

5: Treatment
E 4 T

a , O l
(3 3 - g A 3
v .7 4
(D

'8 i v 5
8- 2 T' I 6
U)

(5

z

 

1 1 1 1
2002 2003 2004
Year

 

Figure 4.3: Number of weed species from 2002 through 2004 in crop diversity
treatments at the KBS Biodiversity Experiment Plots. Symbols refer to crop diversity
treatments (see Table 4.1). Symbols that are shaded indicate treatments with cover
crops. Values are means (:t SE; n =12) averaged across the three main crops.

84

 

   

 

 

5‘ Treatment
3 4 — —- O 1
E C) o 2
G d. 1
9?, ~17 e 3
TD“ 3 __ y 7 4
.g v 5
3 i - 6
(I)
2 2 —

1 1 1 1 1

2002 2003 2004
Year

Figure 4.4: Number of weed species from 2002 through 2004 in cropping-system
diversity treatments planted to winter wheat. Symbols refer to crop diversity treatments
(see Table 4.1). Symbols that are shaded indicate treatments with cover crops. Values
are means (i SE; n = 4).

85

and 4 suggesting that presence of a cover crop, rather than crop rotation, was the
important determinant of weed species richness in wheat.

The impact of crop diversity on weed diversity varied, depending on what
metric of weed diversity was analyzed. Species evenness was affected by crop identity
and year of sampling (crop effect: F251 = 17.18, P < 0.0001; year effect: Fuog = 4.92, P
= 0.009), but not crop diversity, and was lower in soybean compared to corn and wheat.
In contrast, analysis of the Shannon and Simpson’s diversity indices revealed a three-
way interaction between crop diversity, crop identity, and year of sampling (interaction,
S: Fmog = 2.0, P = 0.006; D: F29,109 = 1.86, P = 0.012). Separate repeated measures
analyses for each crop revealed that diversity treatment effects on both diversity
measures were signiﬁcant only in wheat and only in 2004 (interaction, H: F933 = 6.9, P
< 0.001; D: F933 = 5.88, P < 0.0001). In general, weed diversity in wheat was lower in
the three-crop rotations with cover crops than the three-crop rotation without cover
crops (Tukey HSD Test; Figure 4.5).

Community composition and structure

A total of 54 different weed species were observed over the three years of this
experiment; of these 34 were observed in com, 31 in soybean, and 47 in wheat. The
importance of the identity of the main crop on weed species composition was apparent
in the NMDS ordination of presence/absence data, which showed that weed
communities in wheat were distinct ﬁom those observed in corn and soybean (Figure
4.6). ISA indicated that species that were signiﬁcantly associated with winter wheat

included broadleaf and grass species such as prostrate knotweed (Polygonum aviculare

86

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

20- 0.8— b
ab
A a 0.7- B
b
_ ab + _ a
ab ab + O.5~ ac
I 1.0- D 0.41-
%. bc i
0.3~
c
0.5- 0.2-
c
0.1 —
0.0 0.0
1 2 3 4 5 6 1 2 3 4 5 6
Treatment Treatment

Figure 4.5: Shannon diversity (A) and Simpson’s index of species diversity (B) of the
weed community in winter wheat in 2004 in six cropping-system diversity treatments at
the Biodiversity Experiment Plots. Means (:1: SE; n = 4) followed by the same letter are
not signiﬁcantly different at the P < 0.05 level of probability (Tukey HSD test). See
Table 4.1 for treatment codes.

87

 

 

 

 

 

 

 

Crop
I
1.5" 9K A com
I I
9K I soybean
I l 9K 95K
9K wh t
-. 9K 9K ea
)K
N 05‘ 3%
.22
2 I
I
-0.5-1
1
-1.5<
{a r {o oTo 17

Figure 4.6: Plot scores for the ﬁrst two most explanatory axes of non-metric
multidimensional scaling analysis of weed composition (presence/absence) from 2002
through 2004 in six cropping-system diversity treatments (Table 4.1) at the Biodiversity
Experiment Plots at the Long Term Ecological Research site at W. K. Kellogg

Biological Station in Hickory Corners, MI. NMDS: stress = 17.992, P < 0.05,
cumulative r2 for axes 1 and 2 = 0.638.

88

L.) and fall panicum (Panicum dichotomzﬂorum Michx.). Species associated with corn
and soybean included broadleaf species such as common chickweed(Ste11aria media
(L.) Vill.) and common lambsquarters (Chenopodium album L.). Overall, more weed
species were associated with winter wheat than corn or soybean (ISA, Table 4.2).

In addition to the strong effects of crop type on weed communities, we were also
able to detect signiﬁcant crop diversity effects on weed community composition and
abundance. The distance-based multivariate analysis indicated that for each crop, weed
species composition and abundance was affected by crop diversity, year of sampling,
and their interaction (Table 4.3). NMDS ordination of the species abundance data from
2004 showed that within each main crop, treatment replicates with higher crop diversity
tended to ordinate separately from replicates with lower crop diversity (Figure 4.7).
This was particularly evident in wheat; treatments without cover crops clustered
together and were distinct from those with cover crops, which were fairly
homogeneously dispersed in ordination space (Figure 4.70).

Differences among treatments were reﬂected in the ISA analysis on data from
each main crop in 2004. In com the ISA analysis indicated that Canada bluegrass (Poa
compressa L.), was most associated with treatment 3 (two-crop rotation; ISA: IV =
53.6, P = 0.028). For soybean, three species showed strong associations with particular
diversity treatments. Dandelion (T araxacum oﬂicinale G.H. Weber ex Wiggers) and
large crabgrass (Digitaria sanguinalis (L.) Sc0p.) were both strongly associated with
the soybean phase of the three-crop rotation without cover crops (treatment 4) (ISA,
dandelion IV = 70.8, P = 0.003; crabgrass: IV = 53.3, P = 0.042) and giant foxtail

(Setariafaberi Herrm.) was associated with soybean grown in continuous monoculture

89

Table 4.2: Results of indicator species analysis (ISA) on weed community composition
(presence/absence) at the Biodiversity Experiment Plots. ISA Indicator Value (IV)
calculated from annual biomass harvest averaged over three years (2002-2004). Bold
type indicates signiﬁcant associations with a crop species at the P < 0.05 level of
probability.

 

 

 

Species ISA

IVf P—value CropI
Abutilon theophrasti Medik. 22.8 0.002 corn
Amaranthus retroﬂexus L. 4.0 0.688 corn
Ambrosia artemisiifolia L. 39.2 0.001 wheat
Arabidopsis thaliana (L.) Heynh. 30.3 0.001 wheat
Cardamine spp. 13.7 0.008 soybean
Chenopodium album L. 36.1 0.002 corn
Conyza canadensis (L.) Cronq. 33.0 0.001 wheat
Daucus carota L. 5.5 0.109 corn

Digitaria ischaemum (Schreb.) Schreb. ex 33.8 0.001 wheat
Muhl.

Digitaria sanguinalis (L.) Scop. 34.3 0.002 wheat
Elymus repens (L.) Gould 19.5 0.001 wheat
Lamium spp. 24.9 0.001 corn
Mollugo verticillata L. 12.9 0.049 corn
Oxalis stricta L. 45.5 0.001 wheat
Panicum dichotomiﬂorum Michx. 45.1 0.001 wheat
Poa compressa L. 14.9 0.358 soybean
Polygonum aviculare L. 53.7 0.001 wheat
Polygonum convolvulus L. 14.6 0.001 wheat
Polygonum persicaria L. 12.2 0.001 wheat
Portulaca oleracea L. 8.2 0.085 corn
Rumex obtustfolius L. 8.2 0.09 wheat
Setariafaberi Herrm. 18.5 0.006 wheat
Setaria pumila (Poir.) Roemer & J.A. 17.6 0.003 wheat
Schultes

Setaria viridis (L.) Beauv. 13.7 0.002 wheat
Solanum ptychanthum Dunal 11.6 0.004 corn
Sonchus oleraceus L. 7.2 0.033 corn
Stellaria media (L.) Vill. 40.1 0.001 corn
T araxacum ofﬁcinale G.H. Weber ex 21.2 0.085 corn
Wiggers

Veronica gregrina L. 6.8 0.04 wheat

 

* indicator value (% perfect indication).
* crop in which species association was highest.

90

Table 4.3: Results of distance-based multivariate analysis of variance (Anderson 2001,
McArdle and Anderson 2001) of the effects of blocking (B), crop diversity (D), and
year (Y) on weed abundance and composition from 2002 through 2004 in the KBS
Biodiversity Experiment Plots.

 

 

 

Crop Factor df SS MS pseudo- P-value
F
Perm“ Monte
CarloI
Corn B 3 23971.37 7990.46 2.84 0.0004 0.0001
D 5 20795.06 4159.01 1.48 0.0442 0.0516
ErrorM 15 42216.94 2814.46
Y 2 29699.24 14849.62 5.34 0.0001 0.0001
D x Y 10 36023.64 3602.36 1.30 0.0553 0.0611
Errors 36 100021.56 2778.38
Total 71 252727.81
Soybean B 3 25079.54 8359.85 2.53 0.00010 0.00020
D 5 26866.32 5373.26 1.63 0.00700 0.01000
ErrorM 15 49562.44 3304.16
Y 2 23630.92 11815.46 3.32 0.00010 0.00010
D x Y 10 42695.66 4269.57 1.20 0.08350 0.10210
Errors 40 142508.75 3562.72
Total 75 310225.27
Wheat B 3 13986.91 4662.30 1.57 0.01890 0.03170
D 5 48695.94 9739.19 3.28 0.00010 0.00010
ErrorM 15 44524.37 2968.29
Y 2 17731.92 8865.96 4.09 0.00010 0.00010
D x Y 10 35796.18 3579.62 1.65 0.00010 0.00050
Errors 32 69423.52 2 169.49
Total 67 225207.75

 

* 9999 permutations.

:P-value based on Monte Carlo samples from the asymptotic permutation distribution.

91

a2; .0

W

 

m6.

. md

m...

A 2.6 n N.— .mod v & .mm.w~ .1. $05 ”323»
mmwmd M NH .36 v & .S.£ u $05 “$218 636 n N.— .mod v & {hf n 805. “Eco .mezz dz $62.80
boo—oi E cozmam Hammo—omm mwo=oM .M .3 “a 86 58.88% Fame—com ESP wcoq 2: an voom 5 CV “was?

5:33 25 Amv seesaw A3 58 E 26 033» £5.53: 3?ch 5m 3 556888 use. oogwcsnm @003 Mo

max—mam mew—mom 38656622: oEoEéo: mo moxm b03533 $08 95 SE 05 com 888 BE “be 235

a I a D a D a 4 a. x O ”3858:

 

Fm_x<
m? md mQ me.
. . . . . . -.m.v
0 <00 .
4 O i no
I I .
Imo DD
b D fwd
>D
I
be
>
cmogomm .

 

 

mF m6 mo. m6
i . . > . . smé-
I O Q .
DQO >
OD.
D . md
I I I .
. m4
I >
500 .< .

 

92

(treatment 1; ISA: IV = 47.7, P = 0.023). Indicator species analysis identiﬁed ﬁve crop
diversity treatment associations in wheat: common lambsquarters, yellow woodsorrel
(Oxalis stricta L.), and dandelion where all associated with the wheat phase in treatment
4 (three-crop rotation without cover crops; ISP: lambsquarters: IV = 84.0, P = 0.001;
woodsorrel: IV = 70.3, P = 0.001; dandelion: IV = 49.0, P = 0.006). Quackgrass
(Elymus repens (L.) Gould) was associated with treatment 3 (two-crop rotation; IV =
61.3, P = 0.029) and smooth crabgrass Digitaria ischaemum (Schreb.) Schreb. ex
Muhl., was associated with treatment 1 (continuous monoculture; IV = 51.2, P = 0.001).
Soil and light resources and relationships to the weed community

In 2002, two years after the treatments had been initiated, there were indications
of treatment effects on soil resources among crops; however, these were variable and
depended on the identity of the main crop (ANOVA, diversity—by-crop identity
interaction: F1051 = 2.91 , P = 0.006). Across replicates that had been sown to winter
wheat in fall 2001, soil N measured in November 2002 was signiﬁcantly higher in
treatment 2 (monoculture with cover crops) compared to treatments 3, 4, and 5 (Tukey
HSD Test, P < 0.05). However, there were no differences in soil N among plots sown
to corn and soybean in 2002. Treatment effects were also not apparent among plots that
were sown to wheat in October 2002 (a month prior to sampling) (P = 0.093), or those
plots that would be sown to corn in the spring of 2003 (P = 0.61). Plots that would be
sown to soybean in the spring of 2003 did show treatment differences (F5,19 = 3.87, P =
0.013); soil N pools and net mineralization rates were highest in the continuous
monoculture (treatment 1) compared to the other diversity systems (Tukey HSD test, P

< 0.05).

93

Two years later (4 yrs after treatments were initiated), treatment effects were
even more pronounced, but still variable among main crops. Soil N pools measured in
June 2004 showed a signiﬁcant effect of crop diversity in corn and wheat, but not in
soybean (interaction: F1051 = 2.98, P = 0.005) (Figure 4.8). In August, crop diversity
treatment effects on soil N pools were still signiﬁcant for corn and wheat crops. In
corn, soil N levels were higher in treatment 5 than treatments 1, 3, and 4; in wheat, N
levels were higher in treatment 2 than treatment 3 and 4 (Tukey HSD test P < 0.05).

By November, soil N pools in plots that had been sown to soybean in 2004 and wheat in
2003 varied across diversity treatments (interaction: F1051 = 5.21, P < 0.0001), likely
reflecting the recent planting to wheat (in October) in some of these treatments. Similar
to 2002, soil N levels in plots sown to wheat in 2003 were higher in treatment 2 than
treatments 3, 4, 5, and 6. Net soil N—mineralization also varied among cropping-system
diversity treatments that were sown to wheat in 2003, but not those sown to corn and
soybean in 2004 (interaction: F105. = 2.89, P = 0.0062) (Figure 4.9). However, unlike
in 2002, there were signiﬁcant treatment differences on N mineralization among plots
sown to wheat in October 2004 (F 5,15 = 5.91, P = 0.003), with treatment 2 having higher
N pools than treatments 3, 4, and 6 (Tukey HSD test, P < 0.05). This may reﬂect the
relative importance of cover crops versus soybean as a source of N-ﬁxation, since
treatments 3, 4, and 6 would each have been planted to soybean previously (May 2004)
and where cover crop biomass tended to be low compared to wheat (see discussion).

Soil moisture levels measured in late June 2004 were unaffected by crop
diversity (P = 0.0708), but were affected by crop identity (F251 = 23.58, P < 0.0001) and

were lowest in wheat. At late-season (late August), soil moisture levels differed among

94

A. com 2004 B. wheat 2004

 

 

 

 

 

 

 

 

24— 5—
b A c

a a
a a 4-
'v 15— I
I
z ab f ac
3, ab ab 5° 3- —1— ab
2 + a E 3" ab
V 2
Z 8'" .-
’3 a 5 _ b
a) co 2

0 , 1

1 2 3 4 5 6 1 2 3 4 5 6
Treatment Treatment

Figure 4.8: Soil inorganic N pools (NO3-N + NH4-N) in treatments planted to (A) com
and (B) winter wheat in six cropping-system diversity treatments (see Table 4.1) at the
KBS Biodiversity Experiment Plots. Values are means (:t SE; n = 4); bars with the
same letter are not signiﬁcantly different at the P < 0.05 level of probability (Tukey
HSD test). Note different x-axis scales in panels A and B.

95

0.8 -

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3: 0.7 —
g a
CD
3) 0.6 r—
3-
s 05— ab
‘66 ‘1 ab
.5
E 0.4 —
g b
2 - 1
0.2 —- i
0.1
1 3 4
Treatment

Figure 4.9: Net N-mineralization (ugN/g dry soil/day) in wheat in six cropping-system
diversity treatments (Table 4.1) at the Biodiversity Experiment Plots. N-mineralization
rate was measured following 28-day incubation (25 0C) of soils collected on 16
November 2004. Values are means (:1: SE; n = 4); bars followed by the same letter are
not signiﬁcantly different at the P < 0.05 level of probability (Tukey HSD test).

96

crops (F251 = 13.33, P < 0.0001) and crop diversity treatments (F551 = 6.66, P < 0.0001)
and were higher in corn compared to soybean and wheat (Tukey HSD test, P < 0.05).
Overall soil moisture in August tended to be lower in the higher diversity treatments
(Figure 4.10). By 16 November there were no signiﬁcant differences in soil moisture
related to crop type or crop diversity treatment (data not shown).

Light levels (% PAR) at ground level in late July differed among diversity
treatments in wheat but not com or soybean (interaction: F10,42 = 2.14, P = 0.0426) and
were lowest in the highest diversity wheat treatments, particularly those with cover
crops. At the end of the growing season (22 August) light levels differed among crop
diversity treatments in corn and wheat but not in soybean (interaction: F1032 = 5.32, P <
0.0001). In corn light levels were highest in treatment 1 (continuous monoculture). In
wheat they were highest in the two- and three—crop rotation without cover crops. By 19
September light levels differed among diversity treatments in wheat, but not in corn and
soybean (interaction: F10,“ = 32.82, P < 0.0001); wheat treatments with cover crops had
signiﬁcantly lower light levels compared to those without cover crops (Figure 4.11).

The relationship between resource levels and weed community structure in 2004
were investigated by examining correlations between resource variables measured in
2004 and the NMDS ordination axes scores. Data on cover crop biomass was also
included in these analyses. There was little correlation between resource levels in 2004
and the NMDS ordination axes scores in soybean (highest 1'2 = 0.169). In corn, only one
variable, late-season soil moisture had an r2 greater than 0.2 (12 = 0.214). In contrast,

ﬁve variables showed relatively strong correlations (r2 >0.2) with the ordination axes

97

20.-

 

 

 

a
§15— l
ab
CI?» 1 abc 1
== 1
63 1 1
1o—
5

 

 

 

 

 

 

 

 

 

 

 

1 2 3 4
Treatment

 

Figure 4.10: Soil moisture levels in relation to crop diversity treatments (see Table 4.1)
at the KBS Biodiversity Experiment Plots. Soils were sampled at the end of the
growing season 4 yrs after the treatments were initiated (August 2004). Bars are means
(:t SE; n = 12) are averaged across the three crops; bars with the same letter are not
signiﬁcantly different at the P < 0.05 level of probability (Tukey HSD test).

98

 

 

 

 

 

 

 

100— a +
r-I— —I—-
80*
’E
:1
(D
3 60—
o\°
CE __
< 40
D.
b
20:- b b
, I 1L

 

1 2 3 4 5 6
Treatment

Figure 4.11: Light levels in winter wheat grown in different crop diversity treatments
on 19 September 2004. Means (:t SE; n = 4) followed by the same letter are not
signiﬁcantly different at the P < 0.05 level of probability (Tukey HSD test).

99

Table 4.4: Pearson correlation coefﬁcients and 1'2 between environmental and biological
variables and axes scores from NMDS ordination analysis of weed abundance and
composition 2004 in the KBS Biodiversity Experiment Plots.

 

 

 

 

Crop Variablel Axis 1 Axis 2
r r2 r 11
Corn HZOM 0.462 0.214 -0.049 0.002
Soybean - - - - -
Wheat HZOM 0.491 0.241 -0.217 0.047
HZOL 0.510 0.260 -0.407 0.166
PARM 0.541 0.293 -0.561 0.315
PARp 0.470 0.221 -0.534 0.285
COVERBM -0.444 0.198 0.542 0.294

 

* Eight variables were analyzed, only those with r2 > 0.185 for at least one axis are included in the table.
Subscripts after variables are: M, mid-season; L, late-season; P, post-season; BM, biomass (m2).

100

scores in wheat, the strongest of which were those related to cover crop biomass and
light availability (Table 4.4).

Relationships between the weed community and soil resources and light were
also investigated using forward selection multivariate multiple regression. The analysis
indicated that no resource variables explained any of the variation in community
structure in corn (smallest P—value = 0.09) (data not shown). In soybean, light level
(PAR) late in the season was identiﬁed as being a signiﬁcant predictor of weed
community structure (r2 = 0.08, P = 0.047). When forward selection was applied, two
factors, late season light, and mid-season soil moisture were identiﬁed as signiﬁcant,
and together explained 17 % of the variance in community structure in soybean (Table
4.5). Grouping of the resource variables did not improve the strength of the relationship
(data not shown). In contrast to corn and soybean, soil moisture, light, and cover crop
biomass were all signiﬁcant predictors of weed community structure in wheat (data not
shown). Forward selection of the variables indicated that post-season light levels and
N mineralization rate together explained 24 % of the variation in weed community
structure (Table 4.5). Grouping the factors improved the strength of the relationship,
with each group (water, N, light, and cover crop) separately explaining over 12 % of the
variance in community structure. However, forward selection of the model with the
variables grouped indicated that only light resources were signiﬁcant predictors of
community structure, together explaining 24 % of the variation (P = 0.0001).

DISCUSSION
The results of this study suggest that the impacts of crop diversity on weed

communities can be highly variable and depend on the identity of the main crop and the

101

Table 4.5: Results of distance-based multivariate multiple regression with forward
selection (McArdle and Anderson 2001) on weed abundance and composition in 2004
in the KBS Biodiversity Experiment Plots.

 

 

 

CropT Variable:c SS Pseudo- P-value Variance
(Trace) F (perm) explained
Prop. Cum.
Prop.
Corn - - - - - -
Soybean PAR 7071.441 1.8966 0.0471 0.0794 0.0794
H20 7824.207 2.2144 0.0130 0.0878 0.1672
Wheat PAR 16542.100 4.9417 0.0001 0.1834 0.1834
Nmm 5529.253 1.7047 0.0176 0.0613 0.2447

 

I For each crop, order of variables is in decreasing proportion of variance explained.

tEight variables were included in the analysis, only those variables that were signiﬁcant at the P < 0.08
level of probability (9999 permutations) are included in the table.

102

number of crOps in a rotation. Our results also show that increasing crop diversity can
have signiﬁcant impacts on the availability of light and soil resources within the
cropping system, and that these impacts vary depending on the identity of the main
crop. In wheat, differences in weed abundance and diversity among crOp diversity
treatments appeared driven mainly by the presence of cover crops and their impacts on
the availability of light. However, in corn and soybean, there were few effects of crop
diversity on any attribute of the weed community, despite signiﬁcant differences in
resource levels among treatments. In contrast to the effects of cover crops, the rotation
of main crops had relatively little effect on weed communities.

The importance of cover crops in driving the crop diversity-ecosystem function
relationships observed in this study is further supported by the observation that there
were more differences among weed communities in crop diversity treatments in wheat
compared to corn and soybean. The suppressive effects of cover crops have been
shown to vary with cover crops biomass (Akemo et al., 2000; Barberi and Mazzoncini,
2001), which was much higher in wheat compared to corn and soybean. By the time
weeds were harvested (August-September) cover crop biomass in wheat in treatments 2,
5, and 6 averaged 159, 254, and 277 g m'z, respectively, compared to the same
treatments in corn and soybean which each averaged less 12 g m’z. Differences in cover
crop biomass between crops was likely due to the earlier planting of cover crops in
wheat (March) compared to corn and soybean (July).

The impacts of cover crops on weed abundance and resource levels in this study
are consistent with observations from other agronomic studies showing that

management systems that include cover crops often have lower weed abundance and

103

higher soil fertility (Teasdale 1996, Mutch et al. 2003, Drinkwater et al. 1998,
Drinkwater et al. 2000, Snapp et al. 2005). Because there were no fertilizers or
herbicides used in this study, the effects of cover crops on weeds and resource levels in
wheat can be attributed to the cover crops themselves rather than associated
management practices (Doucet et al. 1999, Drinkwater et al. 2000). The weed
suppressive effects of cover crops have been shown to be weed species-speciﬁc
(Creamer et al., 1996; Moonen & Barberi, 2004), perhaps explaining the reduced
species richness and diversity in treatments planted with cover crops (treatments 2, 5,
and 6; Figures 4.3-4.6), and the strong associations between certain weed species and
the diversity treatments. The three cover crop species used in this study have all been
shown to have allelopathic properties (Creamer et al. 1996, Davis and Liebman 2003);
though it is unclear whether, and to what extent allelopathy, rather than competitive
interactions, was responsible for the observed effects on weed suppression in this study.
The impact of cover crops on resource levels and competitive outcomes
observed in our study are also consistent with observations from diversity studies
performed in grassland systems that report strong species-speciﬁc effects on ecosystem
functions (Tilman et al. 1997, Schlapfer and Schmid 1999, Lambers et al. 2004, Hooper
et al. 2005). In some of these studies leguminous forbs have been shown to contribute
to substantial over-yielding in other species through their effects on soil N availability
(Lambers et al. 2004). In contrast, there was no evidence in this study that weeds
beneﬁted from the increased N provided by the legume cover crops or rotation with

soybean, as there were no differences in weed biomass among crop diversity treatments

104

in corn and soybean, and differences in weed abundance among treatments in wheat
appeared due to suppression rather than facilitation.

There is evidence in the ecological literature that experimental communities
containing more species are more resistant to colonization and establishment by new
species (F argione and Tilman 2005), and it is hypothesized that the suppression of
colonization is due to species-rich communities having fewer unexploited niches
(Levine and D’Antonio 1999). While our study did not manipulate the number of crop
species in a single year beyond one main crop and two cover crop species (i.e. treatment
6), the presence of cover crops in wheat certainly led to reduced resource availability—
light levels at ground level were lower in these treatments, which likely led to the
reduced species abundance and lower species diversity at higher levels of crop
diversity. Reduced resource availability due to cover crops also may have reduced
colonization by weed species from surrounding areas or ﬁom the soil seed bank
(Kitajima and Tilman 1996).

Rotation of crops (temporal diversity) appeared to have relatively little effect on
the abundance and composition of the weed community. The relative lack of a rotation
effect was somewhat surprising given that many agronomic studies have reported
signiﬁcant effects of rotating crops on weed abundance and diversity (Liebman and
Dyck 1993, Kegode et al. 1999, Cardina et al. 2002, Legere and Samson 2004,
Westerman et al. 2005). However these results are not without precedent. A number of
recent studies have reported little or no effect of crop rotation on weed communities
(Doucet et al. 1999, Barberi and Lo Cascio 2001), particularly when the effects of

associated management practices are taken into account. For instance, Doucet et al.

105

(1999) reported that after nine years of a com-soybean-wheat rotation following
planting to corn in the tenth year weed communities were similar to continuous corn.
The lack of associated herbicide applications in our system may have decreased the
diversity and strength of stress and mortality factors that weeds were subjected to, and
that would have otherwise been present, in diversiﬁed crop rotations under chemical
management (Liebman and Gallandt 1997, Westerman et al. 2005). The lack of a
consistent rotation effect may also indicate that rotation alone is not a signiﬁcantly
strong ﬁlter on the assembly of weed communities (Booth and Swanton 2002, Cardina
et a1. 2002).

Differences in weed community composition among crops were likely due, in
part, to the timing of seedbed disturbance associated with primary tillage, and variation
in germination and emergence times among weed species (Keever, 1979, Crawley,
2004). Other studies have noted similar differences between the weedy ﬂora in fall and
spring-sown crop species (Andersson and Milberg 1998; Hald 1999b; Hallgren et al.
1999; Squire et a1. 2000). Additional evidence for the importance of the timing of
seedbed disturbance comes from the observation that weeds in corn and soybean tend to
be similar to those in spring-tilled fallows, while weeds in winter wheat tend to be
similar to those in fall-tilled fallows (Appendix A4.l).

The design of the experiment did not allow us to test all possible combinations
of crops or crop diversity—for instance, there was no cover crop monoculture or com-
wheat rotation. Because growers are not likely to grow cover crops alone (without an
associated crop), or rotate corn and wheat in succession; such treatments would not

reﬂect potentially realistic cropping situations. Therefore our conclusions regarding the

106

importance of crop diversity per se are necessarily constrained by the desire to maintain
realistic cropping sequences. However, within this constraint, increasing crop diversity
per se appeared to have few signiﬁcant impacts on the abundance, diversity, and
composition of the weed community. However, because this is a long-tenn study,
treatment effects may develop over time.

The observation that cover crops, rather than crop rotation or diversity per se
appeared to be the main driver of weed abundance does not negate the importance of
increasing crop diversity as a weed and crop management tool (Westerman et al. 2005),
particularly given the observed effects of crop diversity on soil resource levels. Results
of a study examining the response of grain yields to increasing crop diversity showed
that yields in diverse corn and soybean rotations were maintained at levels comparable
to those in conventionally managed systems, despite having higher weed abundance
(Smith Chapter 5). Together, these two studies suggest that increasing crop diversity
may reduce the need for intensive weed management, not because weed abundance is
reduced, but because increased resource availability decreases the impacts of crop-weed
competition (Blackshaw et al. 2002).

ACKNOWLEDGMENTS

We acknowledge and thank the crew of the KBS LTER project, particularly J.
Simmons (farm manager) for their assistance with this project. C. Baker, L. Bell, J.
Dembs, R. Disney, M. Gorentz, M. Hammond, C. Hayes, J. Hernandez, S.

J anuchowski, B. Jasiak, M. Mittelbach, J. Mittelbach, J. Morrison, P. Moseley, N.
Rathbun, J. Rensch, K. Sales, C. Start, L. Sutkowi, C.J. Wyckoff, P. Wyckoff, M.

Yancey, and B. Young assisted with the ﬁeld sampling and lab and greenhouse work.

107

Support for this study was provided by the NSF funded KBS LTER project (DEB98-

10220) and the Mott Program in Sustainable Agriculture (RGS).

108

LITERATURE CITED

Akemo, M.C., Regnier, E.E., Bennett, MA. 2000. Weed suppression in spring-sown
rye (Secale cereale)-pea (Pisum sativum) cover crop mixes. Weed Technology
14: 545-549.

Altieri, MA. 1999. The ecological role of biodiversity in agroecosystems. Agriculture
Ecosystems and Environment 74: 19-31.

Anderson, M. J. 2001. A new method for non-parametric multivariate analysis of
variance. Austral Ecology 26: 32-46.

Anderson, M.J. 2003. DISTLM forward: a FORTRAN computer program to calculate a
distance-based multivariate analysis for a linear model using forward selection.
Department of Statistics, University of Auckland, New Zealand.

Anderson, M.J. 2004. DISTLM v.5: a FORTRAN computer program to calculate a
distance-based multivariate analysis for a linear model. Department of Statistics,
University of Auckland, New Zealand.

Andersson, T.N., Milberg, P. 1998. Weed ﬂora and the relative importance of site,
crop, crop rotation, and nitrogen. Weed Science 46: 30-38.

Barberi, P., Lo Cascio, B. 2001. Long-terrn tillage and crop rotation effects on weed‘
seedbank size and composition. Weed Research 41: 325-340.

Barberi, P., Mazzoncini, M. 2001. Changes in weed community composition as
inﬂuenced by cover crop and management system in continuous corn. Weed
Science 49: 491-499.

Blackshaw, R.E., Semach, G., J anzen, H.H. 2002. Fertilizer application method affects
nitrogen uptake in weeds and wheat. Weed Science 50: 634-641.

Booth, B.D., Swanton, C.J. 2002. Assembly theory applied to weed communities. Weed
Science 50: 2-13. ’

Cardina, J ., Herms, C.P., Doohan, DJ. 2002. Crop rotation and tillage system effects on
weed seedbanks. Weed Science 50: 448-460.

Cardinale, B.J., Palmer, MA. 2002. Disturbance moderates biodiversity-ecosystem
function relationships: Experimental evidence from caddisﬂies in stream

mesocosms. Ecology 83: 1915-1927.

Cottingham, K.L., Brown, B.L., Lennon, J.T. 2001. Biodiversity may regulate the
temporal variability of ecological systems. Ecology Letters 4: 72-85.

109

Cousens, R. 1985. A simple model relating yield loss to weed density. Annals of
Applied Biology 107: 239-252.

Crawley, M.J. 2004. Timing of disturbance and coexistence in a species-rich ruderal
plant community. Ecology 85: 3277-3288.

Creamer, N.G., Bennett, M.A., Stinner, B.R., Cardina, J ., Regnier, BE. 1996.
Mechanisms of weed suppression in cover crop-based production systems.
HortScience 31: 410-413.

Davis, A.S., Liebman, M. 2003. Cropping system effects on giant foxtail (Setaria
faberi) demography: 1. Green manure and tillage timing. Weed Science 51: 919-
929.

Dekker, J. 1997. Weed diversity and weed management. Weed Science 45: 357-363.

Doucet, C., Weaver, S.E., Hamill, A.S., Zhang, J. 1999. Separating the effects of crop
rotation from weed management on weed density and diversity. Weed Science
47: 729-735.

Drinkwater, L.E., Wagoner, P., Sarrantonio, M. 1998. Legume-based cropping systems
have reduced carbon and nitrogen losses. Nature 396: 262-265.

Drinkwater, L.E., Janke, R.R., Rossoni-Longnecker, L. 2000. Effects of tillage intensity
on nitrogen dynamics and productivity in legume-based grain systems. Plant
and Soil 227: 99-113.

Dufrene, M., Legendre, P. 1997. Species assemblages and indicator species: the need
for a ﬂexible asymmetrical approach. Ecological Monographs 67: 345-366.

F argione, J .E., Tilman, D. 2005. Diversity decreases invasion via both sampling and
complementarity effects. Ecology Letters 8: 604-611.

Hald, AB. 1999. The impact of changing the season in which cereals are sown on the
diversity of the weed ﬂora in rotational ﬁelds in Denmark. Journal of Applied
Ecology 36: 24-32.

Hallgren, E., Palmer, M.W., Milberg, P. 1999. Data diving with cross-validation: an
investigation of broad-scale gradients in Swedish weed communities. Journal of
Ecology 87: 1037-1051.

Hector, A., Schmid, B., Beierkuhnlein, C., Caldeira, M.C., Diemer, M.,
Dimitrakopoulos, P.G., Finn, J.A., Freitas, H., Giller, P.S., Good, J ., Harris, R.,
Hogberg, P., Huss-Danell, K., Joshi, J., Jumpponen, A., Komer, C., Leadley,
P.W., Loreau, M., Minns, A., Mulder, C.P.H., O'Donovan, G., Otway, S.J.,
Pereira, J.S., Prinz, A., Read, D.J., Scherer-Lorenzen, M., Schulze, B.D.,

110

Siamantziouras, A.S.D., Spehn, E.M., Terry, A.C., Troumbis, A.Y., Woodward,
F.I., Yachi, S., Lawton, J .H. 1999. Plant diversity and productivity experiments
in European grasslands. Science 286: 1123-1127.

Hooper, D.U., Vitousek, PM. 1998. Effects of plant composition and diversity on
nutrient cycling. Ecological Monographs 68: 121-149.

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., Setala, H., Symstad,
A.J., Vandermeer, J., Wardle, DA. 2005. Effects of biodiversity on ecosystem

functioning: A consensus of current knowledge. Ecological Monographs 75: 3-
35.

Huston, M. 1979. General hypothesis of species-diversity. American Naturalist 113: 81-
101.

Jolliffe, RA. 1997. Are mixed populations of plant species more productive than pure
stands? Oikos 80: 595-602.

Keever, C. 1979. Mechanisms of plant succession on old ﬁelds of Lancaster County,
Pennsylvania. Bulletin of the Torrey Botanical Club 106: 299-308.

Kegode, G.O., Forcella, F., Clay, S. 1999. Inﬂuence of crop rotation, tillage, and
management inputs on weed seed production. Weed Science 47: 175-183.

Kitajima, K., Tilman, D. 1996. Seed banks and seedling establishment on an
experimental productivity gradient. Oikos 76: 381-391.

Lambers, J .H.R., Harpole, W.S., Tilman, D., Knops, J ., Reich, PB. 2004. Mechanisms
responsible for the positive diversity-productivity relationship in Minnesota
grasslands. Ecology Letters 7: 661-668.

Legere, A., Samson, N. 2004. Tillage and weed management effects on weeds in
barley-red clover cropping systems. Weed Science 52: 881-885.

Legere, A., Stevenson, EC. 2002. Residual effects of crop rotation and weed
management on a wheat test crop and weeds. Weed Science 50: 101-111.

Levine, J .M., D'Antonio, GM. 19999. Elton revisited: a review of evidence linking
diversity and invisibility. Oikos 87: 15-26.

Liebman, M., Dyck, E. 1993. Crop rotation and intercropping strategies for weed
management. Ecological Applications 3: 92-122.

Liebman, M., Gallandt, ER. 1997. Ecological management of crop-weed interactions.
In: Ecology in Agriculture. LE Jackson, ed. San Diego, CA, USA. pp 291-343.

111

Liebman, M., Staver, GP. 2001. Crop diversiﬁcation for weed management. In:
Ecological Management of Agricultural Weeds. M Liebman, M Mohler, CL
Staver, eds. Cambridge University Press, Cambridge, UK. pp 322-374.

Little, R. C., G. A. Milliken, W. W. Stroup, and R. D. Wolﬁnger. 1996. SAS system for
mixed models. SAS Institute, Cary, North Carolina, USA.

Loreau, M., Hector, A. 2001. Partitioning selection and complementarity in biodiversity
experiments. Nature 412: 72-76.

Magurran, A.E. 1988. Ecological diversity and its measurement. Princeton University
Press, Princeton, NJ, USA.

McArdle, B.H., Anderson, M.J. 2001. Fitting multivariate models to community data: a
comment on distance-based redundancy analysis. Ecology 82: 290-297.

McCune, B., Grace, J .B. 2002. Analysis of Ecological Communities. MjM Software
Design, Gleneden Beach, USA.

McCune, B., Mefford, M.J. 1999. PC-ORD. Multivariate Analysis of Ecological Data,
Version 4. MjM software Design, Gleneden Beach, Oregon, USA.

Minns, A., Finn, J ., Hector, A., Caldeira, M., Joshi, J., Palmborg, C., Schmid, B.,
Scherer-Lorenzen, M., Spehn, E., Troumbis, A. 2001. The functioning of

European grassland ecosystems: potential beneﬁts of biodiversity to agriculture.
Outlook on Agriculture 30: 179-185.

Moonen, A.C., Barberi, P. 2004. Size and composition of the weed seedbank after 7
years of different cover-crop-maize management systems. Weed Research 44:
163-1 77.

Mutch, D.R., Martin, T.E., Kosola, KR. 2003. Red clover (T rifolium pratense)
suppression of common ragweed (Ambrosia artemisiifolia) in winter wheat
(T riticum aestivum). Weed Technology 17: 181-185.

Pielou, EC. 1969. An introduction to mathematical ecologi. John Wiley and Sons,
New York, USA.

Robertson, G.P., Klingensmith, K.M., Klug, M.J., Paul, E.A., Crum, J.R., Ellis, B.G.
1997. Soil resources, microbial activity, and primary production across an
agricultural ecosystem. Ecological Applications 7: 158-170.

SAS Institute. 1999. SAS/STAT User’s Guide. Eighth edition. Volume 2. SAS Institute,
Cary, North Carolina, USA.

112

Schlapfer, F., Schmid, B. 1999. Ecosystem effects of biodiversity: A classiﬁcation of
hypotheses and exploration of empirical results. Ecological Applications 9: 893-
912.

Snapp, S.S., Swinton, S.M., Labarta, R., Mutch, D., Black, J.R., Leep, R., Nyiraneza, J .,
O'Neil, K. 2005. Evaluating cover crops for beneﬁts, costs and performance
within cropping system niches. Agronomy Journal 97: 322-332.

Squire, G.R., Rodger, S., Wright, G. 2000. Community-scale seedbank response to less
intense rotation and reduced herbicide input at three sites. Annals of Applied
Biology 136: 47-57.

Teasdale, J .R. 1996. Contribution of cover crops to weed management in sustainable
agricultural systems. Journal of Production Agriculture 9: 475-479.

Teasdale, J .R., Mangum, R.W., Radhakrishnan, J., Cavigelli, M.A. 2004. Weed
seedbank dynamics in three organic farming crop rotations. Agronomy Journal
96: 1429-1435.

Tilman, D. 1999. The ecological consequences of changes in biodiversity: A search for
general principles. Ecology 80: 1455-1474.

Tilman, D., Knops, J ., Wedin, D., Reich, P., Ritchie, M., Siemann, E. 1997. The
inﬂuence of functional diversity and composition on ecosystem processes.
Science 277: 1300-1302.

Tilman, D., Reich, P.B., Knops, J ., Wedin, D., Mielke, T., Lehman, C. 2001. Diversity
and productivity in a long-tenn grassland experiment. Science 294: 843-845.

Westerman, P.R., Liebman, M., Menalled, F.D., Heggenstaller, A.H., Hartzler, R.G.,
Dixon, RM. 2005. Are many little hammers effective? Velvetleaf (Abutilon
theophrasti) population dynamics in two- and four-year crop rotation systems.
Weed Science 53: 382-392.

Weiner, J ., Griepentrog, H.W., Kristensen, L. 2001. Suppression of weeds by spring
wheat Triticum aestivum increases with crop density and spatial uniformity.
Journal of Applied Ecology 38: 784-790.

113

CHAPTER 5

CROPPING-SYSTEM DIVERSITY AF FECTS ECOSYSTEM FUNCTION IN
ROW CROPS

ABSTRACT

Understanding the role of diversity in the functioning of ecosystems has
important implications for agriculture. We show that increasing the diversity of a row
crop system, both spatially and temporally by altering crop rotation and cover crops,
increases crop productivity (yield) in the absence of chemical inputs (herbicides,
pesticides, fertilizer). The effects of cropping system diversity were manifest most
strongly in maize and were likely due to the diversity of legumes in the cropping
sequence and their effects on available inorganic soil nitrogen. Weed abundance was
independent of crop diversity and yield. These results demonstrate how ecological
theory is relevant to the management of agricultural ecosystems, and suggests additional
strategies for managing agricultural systems with biological inputs.

INTRODUCTION

While the intensiﬁcation of modern agricultural production systems has led to
increases in world food production (Matson et al. 1997, Tilman et al. 2001), increased
production has come at the expense of cropping system and wild-land diversity and
environmental health (Pimentel et al. 1992, Matson et al. 1997, Green et a1. 2005,
Robertson and Swinton 2005). Conversion of species-rich wild-lands to large-scale
agricultural production, often as monocultures, have reduced the potential for species
interactions important in community and ecosystem-level regulation of pest populations
and nutrient cycling, which must now be replaced by chemical inputs of pesticides and

fertilizer (Drinkwater et al. 1998, Tilman et al. 2001, Robertson and Swinton 2005,

114

Hooper et al. 2005). The environmental impacts associated with such inputs have led
many to question the sustainability of conventional agricultural practices (Pimentel et
al. 1992, Matson et al. 1997, Tilman 1999, Tilman et al. 2001, Reganold et al. 2001,
Green et al. 2005).

The relationship between species diversity and ecosystem function may have
substantive implications for the development of agriculture that is more sustainable
(Matson et al. 1997, Liebman and Davis 2000, Mader et al. 2002, Robertson and
Swinton 2005). Considerable empirical evidence suggests a link between important
ecosystem functions such as nutrient cycling and productivity and the number of species
in a community (Tilman et al. 1996, Mittelbach et al. 2001, Mader et al. 2002, Hooper
et al. 2005). While the particular mechanisms driving the relationship between diversity
and ecosystem function in experimental assemblages are oﬁen debated, they appear to
operate most strongly at relatively low levels of diversity (Huston 1997, Tihnan 1999,
Schwartz et al. 2000, Lambers et al. 2004). Because plant diversity in annual row crop
systems is already low (usually a single crop), theory thus suggests that increased
diversity might be expected to have particularly large effects on ecosystem function in
these systems.

Here we report the results of a study established in 2000 in which cropping
system diversity is manipulated, both temporally and spatially, through the rotation of
both row and cover crops grown without external chemical inputs [of pesticides,
herbicides, or fertilizer]. Crop yields, weed abundance, and soil nitrogen were
measured over the course of the study to determine how crop diversity affects

ecosystem processes related to row-crop production. We grew corn, soybean, and

115

wheat in combinations of 1-3 crop species annually and 1-6 species over a 3-year
rotation period to provide a wide range of plant community diversities.
MATERIALS AND METHODS
Study site

The study was conducted at the Biodiversity Experimental Plots (hereafter BEP,
http://lter.kbs.msu.edu/experimentalDesign.html) at the W. K. Kellogg Biological
Station LTER (Long-term Ecological Research) site of Michigan State University in
SW Michigan, USA. Soils at the study site are a mixture of Kalamazoo (ﬁne-loamy,
mixed, mesic Typic Hapludalfs) and Oshtemo (coarse-loarny, mixed, mesic Typic
Hapludalfs) sandy loams (Sanchez et al. 2001). Annual precipitation at the KBS LTER
site is 890 mm'1 and mean annual temperature is 9.7 °C
(http://lter.kbs.msu.edu/siteDescription.htrnl). 0

Treatment systems were initiated at the BEP in 2000 and consisted of three row
crops, corn (Zea mays L.), soybean (Glycine max (L.) Merr.), and winter wheat
(T riticum aestivum L.) grown in continuous monoculture and in two and three-year
annual rotations with and without cover crops (zero, one, or two legume/small grain
species). The six treatment systems range in diversity from one to three crop species
annually and one to ﬁve species over a three-year period and include: (1) continuous
monoculture (1 spp), (2) continuous monoculture w/ one cover crop annually (2 spp),
(3) two-crop rotation (2 spp), (4) three-crop rotation (3 spp), (5) three-crop rotation with
one cover crop annually (2 spp annually; 4 spp over 3 yr rotation), and (6) three-crop
rotation with two cover crop species annually (3 spp annually; 5 spp over 3 yr rotation).

In the two-crop rotation (system 3), only three of the possible sequences were included

116

in the design: com-soybean, soybean-com, and soybean-wheat. The three-crop
rotations (systems 4-6) are all planted in the same sequence: com-soybean-wheat. Each
phase of the two and three-crop rotations is present every year, allowing comparisons to
be made among the six diversity treatments for each the three main crop species.

Cover crops vary depending on the main crop and include red clover (T rifolium
pratense L.), crimson clover (T rifolium incarnatum L.), and cereal rye (Secale cereale
L.). In treatments receiving a single cover crop (systems 2 and 5), T. pretense is sown
into corn in July and wheat in March, and S. cereale is sown in soybean (system 2 only)
in October. For the highest diversity treatment (system 6) each crop is grown with two
cover crops: T. pratense is planted into corn and wheat as above, and T. incarnatum is
planted into soybean. Secale cereale is added to corn and wheat in October.

Diversity treatments were randomly applied to plots in three linear blocks; in the
fourth block, treatments were arranged in 4 x 5 blocks and planted in sequential order of
decreasing diversity. Individual plots measure 9.1m x 27.4 m with no buffer strips
between plots. Replicate blocks were separated by grass buffers. All plots are chisel
plowed and soil-ﬁnished prior to planting the main crop. Corn and soybean are planted
in late April or May and winter wheat is planted in late September or early October. No
external inputs (herbicides, pesticides, or fertilizer, including manure and compost)
were applied to any of the treatments aﬁer 2001; periodic inter-row cultivations
(cultivation/rotary hoeing) are performed in corn and soybean early in the season to

control weeds.

ll7

Sampling

Crop yield was determined annually using standard practices with a plot
combine. Wheat was harvested in July and soybean and corn was harvested in October
and November, respectively. Crop yields (Kg ha") were calculated assuming 13 %
(wheat and soybean) and 15 % (corn) moisture. Average yields in a nearby
conventionally managed (full inputs of herbicides, pesticides, and fertilizer) com-
soybean-wheat rotation at the KBS LTER were calculated in a similar manner and are
presented in the results to provide a local conventional baseline with which to compare
those in the BEP. Yields for each crop in the conventional system are the average of the
six years that each crop was present in the rotation over the time period from 1989 to
2002.

Aboveground weed biomass was harvested at peak biomass (August-September)
from two 0.25 x 1 m quadrats placed in two central locations in each plot. The location
of sampling was changed each year to minimize edge effects and to avoid harvesting
weeds from areas that had been harvested the previous season. Harvested weed
biomass was separated to species, dried at 65° C for 48 hrs and weighed to the nearest
0.01 g.

Inorganic soil nitrogen (N03' and NIH) was measured at the end of the growing
season (November) in 2002 and at three times (28 June, 23 August, and 16 November)
in 2004. Soils were sampled to a depth of 25 cm (four-2 cm diameter cores) within the
central 3 x 9 m of each plot from both the rows and inter-rows. Aggregated samples
were sieved through a 4 mm screen to remove stones and large pieces of organic

material. Following sieving, a portion of each aggregated sample (20-30 g) was

118

processed for gravimetric soil moisture analysis. The remaining sample was
subdivided; 20 g of each sample was processed for N extraction with 100 m1 of 1M
KC L (20 g) and the rest air-dried and archived. Nitrogen (NO3' and NH4+) was
analyzed with an Alpkem 3550 continuous-ﬂow colorimetric analyzer (OI Analytical,
College Station, TX, USA). Soils were stored at 4° C and were processed within 48 hrs
of collection.
Statistical analyses

Within each crop, separate analyses were performed to determine the effect of
cropping system diversity on yield and aboveground weed biomass across the three
years (2002-2004). To analyze differences among treatments we used a three factor
(diversity system, block, year) repeated-measures ANOVA, with year as the repeated
factor (SAS Proc Mixed procedure) followed by a Tukey-Kramer Test at P = 0.05.
Weed biomass data was log10(X + 0.01) transformed prior to analysis to improve
homoscedasticity. Crop yield (2004) was also analyzed using weed biomass and total
available inorganic soil N as independent variables in separate linear regressions.

RESULTS AND DISCUSSION

Cropping system diversity affected yields in all three crops (Table 5.1).
However, the strongest effects of diversity were observed in corn, where over-yielding
(% increase in yield compared to the monoculture) was over 100% in some treatments.
Yields in the highest diversity corn treatment averaged 6.9 Mg ha'1 (111 bu A")
compared to 3.4 Mg ha‘I (54 bu A") in the lowest diversity treatment. In addition, corn

yields in the most diverse treatments (four and ﬁve species) were higher than those in

119

the conventionally managed treatments on nearby plots, which averaged just under 6
Mg ha’1 (95.5 bu A") over six years from 1989 to 2002 (Figure 5.1).

In soybean and wheat, yield differences among cropping systems were also
signiﬁcant (Table 5.1) but of lower magnitude (32 and 53 %, respectively). Grain
yields in soybean in the highest diversity treatments (systems 4-6) were also higher than
those in the conventionally managed system, which averaged just under 2.2 Mg ha'1 (33
bu A'l) (Figure 5.1). In contrast, yields in wheat at all levels of diversity were
substantially lower than those in the conventional system (3.6 Mg ha"). The signiﬁcant
treatment effect in wheat was due to low yields in the monoculture (system 1); the other
diversity systems did not differ from one another (Tukey-Kramer Test, P > 0.05).

The reasons for the particularly strong diversity effects in corn compared to
soybean and wheat were unclear, but may be due to differences in row-spacing, time of
planting or physiological differences of C4 grasses compared to legumes and C3 species
(Loomis and Connor 1992). C4 grasses also had strong responses to experimental
manipulations of diversity in a grassland ecosystem (Lambers et al. 2004). Because the
responses to cropping system diversity were most pronounced in corn, we focus on this
crop to examine potential mechanisms responsible for the observed diversity effects.

We found no evidence that the observed yield differences were due to reduced
weed-crop competition in these diversity treatments. Weed abundance was independent
of cropping system diversity over the three years for which we had data (P = 0.205)
(Table 5.1); nor was there any relationship between weed biomass and corn yield (r2 =
0.04, P = 0.34). We did not detect any differences in weed abundance in corn among

treatments with and without cover crops, which suggests that cover crops provided few

120

Table 5.1: Repeated measures ANOVA for the effects of cropping system diversity on
crop yields and weed biomass in the KBS Biodiversity Experiment Plots (BEP) over

three years (2002-2004). Signiﬁcant effects are bolded for clarity.

 

 

 

 

Crop yields Weed Biomass

F-value (df) P > F F-value (df) P > F
Corn
System 15.22 (5,15) <0.0001 1.66 (5,15) 0.2054
Year 0.19 (2,36) 0.8294 4.26 (2,36) 0.0219
System*year 2.07 (10, 36) 0.0537 1.48 (10,36) 0.1880
Soybean
System 6.55 (5,15) 0.0020 2.07 (5,15) 0.1269
Year 162.47 (2,40) <0.0001 8.87 (2,40) 0.0006
System*year 1.82 (10,40) 0.0888 1.00 (10,40) 0.4591
Wheat
System 5.01 (5,15) 0.0067 17.81 (5,15) <0.0001
Year 112.15 (2,33) <0.0001 30.44 (2,33) <0.0001
System*year 4.61 (9,33) 0.0005 13.15 (9,33) <0.0001

 

121

 

10.. . corn
9L A soybean
8....
Iwheat
7" t i
0) 6L... --------------------------------------- com
5; 5— 9 §
5 §
; 4“
3:---§ ................ i ---------------- w heat
2 Z.""g""!""i""""'i"':'"' soybean
1...
0 1 1 1 1 1 1 1

 

1 2 3 4 5 6
Diversity system

Figure 5.1: Crop yields averaged across three years (2002—2004) in the BEP
experiment. Values are means (:t 1 SE, n = 12). Dotted lines indicate yields (six year
average) for each crop in a nearby conventionally managed com-soybean-wheat
rotation. The diversity system treatments are: (1) continuous monoculture, (2)
monoculture with one cover crop annually, (3) two crop rotation, (4) three-crop
rotation, (5) three-crop rotation with one cover crop annually, and (6) three-crop
rotation with two cover crops annually.

122

additional weed control beneﬁts beyond those of mechanical cultivation (Knezevic et al.
2002).

In contrast, we found strong evidence for differential facilitation of soil nitrogen
availability across these diversity treatments. Early season soil nitrogen levels in 2004
had a signiﬁcant positive effect on corn yield (r2 = 0.67, P < 0.001) (Figure 5.2), and
was highest in the most diverse systems (systems 4 and 5) and lowest in the least
diverse system (continuous monoculture, system 1) (Tukey-Kramer Test, P < 0.05).
Available soil nitrogen levels remained high in the diverse treatments through the mid-
season sampling (23 August) (F535 = 4.31, P = 0.012) but decreased to levels that were
equivalent across the diversity treatments by the end of the growing season (16
November) (P = 0.1213).

Differences in inorganic nitrogen availability and crop yields among the treatments
appeared to be driven primarily by the diversity of legume species in the cropping
sequence (Drinkwater et a1. 1998). These results parallel those from studies in
grasslands that have attributed diversity effects to the identity and functional traits of
the species (Hooper and Vitousek 1998, Diaz and Cabido 2001, Lambers et al. 2004).

In many of these experiments legumes were identiﬁed as important drivers of positive
diversity-productivity relationships due to their stimulation of over-yielding in non-
nitrogen ﬁxing species (Lambers et al. 2004, Hooper et al. 2005). Corn yields increased
when the number of legume species in the rotation increased from zero to one (system 1
vs. 2) and from one to two (system 5; Figure 5.1). Effects were similar whether the

diversity increase was due to the inclusion of soybean or T rifolium cover crops in the

123

    

 

 

10—
9—
8— V Treatment
0 1
:5 7”
'2 A 2
a 6" 3
z 5* v 4
31.3 V 5
>- 4" I 6
3 r2=0.64.P<0-001
2 7‘?) y=1,344+0.418x
1 ' l I
0 10 20 30

Soil N (NO3 + NH4-pg/g)

Figure 5.2: Relationship between corn grain yields (October) and early season
inorganic soil nitrogen (June) in 2004. Refer to ﬁgure 1 for a description of the
diversity treatments.

124

rotation. Increases in the diversity of non-legume crops did not affect yields
appreciably (i.e. system 2 vs. systems 3 and 4).

Despite the observed signiﬁcant effect of legumes on ecosystem processes in
this experiment, the fact that continuous monocultures did not out-yield the higher
diversity treatments in any of the crops implies that there are beneﬁts to increasing
diversity per se, beyond simply the inclusion of species of large-effect (Lambers et al. .
2004). Other factors may also have contributed to some of the observed effects of
cropping system diversity on yields, such as disease and insect pests (Andow 1991 , Zhu
et al. 2000). Though we did not measure disease or insect damage, we saw no obvious
signs of disease or insect damage among the treatments.

It would be naive to suggest that many growers and agricultural researchers are
unaware of the beneﬁcial effects that crop rotation and cover crops have on crop
performance (Robertson and Swinton 2005). And indeed, many agricultural studies
have been performed to examine these effects (Liebman and Dyck 1993, Drinkwater et
al. 1998). However, such studies are rarely performed within the broader context of
diversity-ecosystem function studies, or not confounded with the application of
fertilizer or pesticide inputs (Doucet et al. 1999).

Results of this study show that increasing diversity in cropping systems can lead
to signiﬁcant increases in ecosystem functions related to crop production, and offers
further supporting evidence that ecosystem functioning is related to both the number
and types of species in a community. In addition, our results ﬁrrther challenge the
conventional notions that high external inputs are always necessary to maintain high

yields and that environmentally friendly farming strategies will necessarily reduce

125

farmer’s proﬁts (Banks 2004). However, while our results suggest that farmers could
reduce their reliance on agricultural inputs by adopting the most diverse cropping
systems studied here, a thorough economic analysis will be necessary to determine
whether or not the reduced expenditures on chemical pesticides and fertilizers offset the
cost of cover crops and the yield reduction during the wheat phase of the rotation.
ACKNOWLEDGMENTS

We wish to thank the KBS LTER farm manager J. Simmons for maintaining the
BEP at the KBS LTER and C. Baker for assistance in sampling and processing these
data. D. Buhler, S. Emery, D. Landis, C. Malmstrom, and G. Mittelbach provided
helpful comments on earlier drafts of this chapter. Support for this study was provided
by the NSF -funded KBS LTER project (DEB98-10220), the CS. Mott Program in

Sustainable Agriculture (ROS), and the Michigan Agricultural Experiment Station.

126

LITERATURE CITED

Andow, D.A. 1991.Vegetational diversity and arthropod population response. Annual
Review of Entomology 36: 561-586.

Banks, J .E. 2004. Divided culture: integrating agriculture and conservation biology.
Frontiers in Ecology and the Environment 2: 537-545.

Diaz, S., Cabido, M. 2001. Vive la difference: plant functional diversity matters to
ecosystem processes. Trends in Ecology & Evolution 16: 646-655.

Doucet, C., Weaver, S.E., Hamill, A.S., Zhang, J .H. 1999. Separating the effects of crop
rotation from weed management on weed density and diversity. Weed Science
47: 729-735.

Drinkwater, L.E., Wagoner, P., Sarrantonio, M. 1998. Legume-based cropping systems
have reduced carbon and nitrogen losses. Nature 396: 262-265.

Green, R.E., Cornell, S.J., Scharlemann, J .P.W., Balmford, A. 2005. Farming and the
fate of wild nature. Science 307: 550-555.

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., Setala, H., Symstad,
A.J., Vandermeer, J ., Wardle, DA. 2005. Effects of biodiversity on ecosystem

functioning: A consensus of current knowledge. Ecological Monographs 75: 3-
35.

Hooper, D.U., Vitousek, PM. 1998. Effects of plant composition and diversity on
nutrient cycling. Ecological Monographs 68: 121-149.

Huston, M.A. 1997. Hidden treatments in ecological experiments: Re-evaluating the
ecosystem function of biodiversity. 0ecologia 110: 449-460.

Knezevic, S.Z., Evans, S.P., Blankenship, E.E., Van Acker, R.C., Lindquist, J .L. 2002.

Critical period for weed control: the concept and data analysis. Weed Science
50: 773-786.

Lambers, J .H.R., Harpole, W.S., Tilman, D., Knops, J ., Reich, PB. 2004. Mechanisms
responsible for the positive diversity-productivity relationship in Minnesota

grasslands. Ecology Letters 7: 661 -668.

Liebman, M., Davis, AS. 2000. Integration of soil, crop and weed management in low-
extemal-input farming systems. Weed Research 40: 27-47.

Liebman, M., Dyck, E. 1993. Crop rotation and intercropping strategies for weed

127

management. Ecological Applications 3: 92-122.

Loomis, R.S., Connor, DJ. 1992. Crop Ecology. Cambridge University Press,
Cambridge.

Mader, P., Fliessbach, A., Dubois, D., Gunst, L., Fried, P., Niggli, U. 2002. Soil fertility
and biodiversity in organic farming. Science 296: 1694-1697.

Matson, P.A., Parton, W.J., Power, A.G., Swift, M.J. 1997. Agricultural intensiﬁcation
and ecosystem properties. Science 277 : 504-509.

Mittelbach, G.G., Steiner, C.F., Scheiner, S.M., Gross, K.L., Reynolds, H.L., Waide,
R.B., Willig, M.R., Dodson, S.I., Gough, L. 2001. What is the observed
relationship between species richness and productivity? Ecology 82: 2381-2396.

Pimentel, D., Acquay, H., Biltonen, M. Rice, P., Silva, M., Nelson, J ., Lipner, V.,
Giordano, S., Horowitz, A., Damore, M. 1992. Environmental and economic
costs of pesticide use. Bioscience 42: 750-760.

Reganold, J.P., Glover, J .D., Andrews, P.K., Hinman, HR. 2001. Sustainability of three
apple production systems. Nature 410: 926-930.

Robertson, G.P., Swinton, SM. 2005. Reconciling agricultural productivity and
environmental integrity: a grand challenge for agriculture. Frontiers in Ecology
and the Environment 3: 38-46.

Sanchez, J .E., Willson, T.C., Kizilkaya, K., Parker, E., Harwood, RR. 2001. Enhancing
the mineralizable nitrogen pool through substrate diversity in long term cropping
systems. Soil Science Society of America Journal 65: 1442-1447.

Schwartz, M.W., Brigham, C.A., Hoeksema, J.D., Lyons, K.G., Mills, M.H., van
Mantgem, RI. 2000. Linking biodiversity to ecosystem function: implications
for conservation ecology. 0ecologia 122: 297-305.

Tilman, D. 1999. Global environmental impacts of agricultural expansion: The need for
sustainable and efﬁcient practices. Proceedings of the National Academy of

Sciences of the United States of America 96: 5995-6000.

Tilman, D., Fargione, J., Wolff, B., D'Antonio, C., Dobson, A., Howarth, R., Schindler,
D., Schlesinger, W.H., Simberloff, D., Swackhamer, D. 2001. Forecasting
agriculturally driven global environmental change. Science 292: 281-284.

Tilman, D., Wedin, D., Knops, J. 1996. Productivity and sustainability inﬂuenced by
biodiversity in grassland ecosystems. Nature 379: 718-720.

Zhu, Y.Y., Chen, H.R., Fan, J.H., Wang, Y.Y., Li, Y., Chen, J.B., Fan, J.X., Yang, SS,

128

Hu, L.P., Leung, H., Mew, T.W., Teng, P.S., Wang, Z.H., Mundt, CC. 2000.
Genetic diversity and disease control in rice. Nature 406: 718-722.

129

CHAPTER 6
CONCLUSIONS AND FUTURE DIRECTIONS

The results from the research presented in this dissertation provide compelling
support for the positive role that increasing cropping-system diversity can play in the
functioning and production of row-crop agricultural systems. We were able to
demonstrate that by manipulating crop rotation and cover crops with complementary
resource use requirements and functional traits there are signiﬁcant positive impacts on
soil resources and crop yields. In contrast to the effects on crop yields, manipulation of
diversity with crops that have complementary resource use requirements and functional
traits appeared to have relatively few impacts on the abundance, composition, and
diversity of the weed community when the impacts of cover crops were taken into
account. Together these two main results suggest that increasing row crop diversity
may reduce the need for intensive weed management, not because weed abundance is
reduced, but because increased resource availability decreases the impacts of crop-weed
competition.

There are relatively few studies that have looked at the role of diversity in
agricultural systems and these too have generally shown beneﬁcial effects of increasing
agroecosystem diversity on soil fertility and crop productivity (Vandermeer 1989,
Drinkwater et al. 1998, Altieri 1999, Reganold et al. 2001,Westerman et al. 2005). Our
results are particularly compelling because we constrained the crop diversity treatments
in both the LFL and the BEP studies to reﬂect potentially realistic cropping situations.

In addition, in the BEP study, we varied only biological diversity and not other

130

components of cropping system management that could potentially be used by growers
to further increase crop ﬁelds and/or fertility in these systems.

The results of this dissertation help ﬁrrther our understanding of the factors that
determine the abundance and composition of weed communities and inﬂuence
ecosystem processes in row-crop agricultural systems. Results of this dissertation also
demonstrate the potential for crop diversity to be used as a management tool and
suggest that growers may have viable alternatives to the use of synthetic chemical
inputs for weed and crop management.

Future directions

More work will be needed to determine the mechanisms underlying the crop
yield responses to increasing diversity observed in this study. This will likely require a
detailed study of the soil microbial community to determine whether more diverse
cropping systems have a greater number/and or diversity of beneﬁcial soil
microorganisms and/or are more resistant to soil pathogen attack (Zak et al. 2003).
Determining the relationship between crop diversity and insect pest abundance may also
yield clues to the importance of crop diversity in this system; results from past studies
comparing numbers of herbivore pests in simple and diverse agroecosystems have been
equivocal (Andow 1991, Tonhasca and Byme 1994, Norris and Kogan 2000).

Future studies of diversity in agricultural systems should address the role of
disturbance in mediating diversity-ecosystem ﬁmction relationships (Huston 1979,
Cardinale and Palmer 2002). While management-related disturbances were minimized
in the BEP study (i.e. no herbicide applications), tillage and inter-row cultivation were

both used to prepare the seed bed and manage weeds early in the growing season. Both

131

of these factors can have signiﬁcant impacts on weed communities (Menalled et al.
2001, Davis et al. 2005) and may have contributed to the lack of response to crop
diversity we observed in the weed community in both the LFL and BEP studies.

No till-cropping systems may be a more appropriate agricultural system within
which to test general theories regarding diversity and ecosystem function because the
lack of annual disturbance under no-till cropping may increase the potential for
competitive and facilitative interactions among crops and weeds. No-till systems tend
also to require more herbicide inputs than tilled systems because weeds cannot be
managed mechanically (Donald et al. 2001). Therefore manipulations of diversity that
decrease the need for weed management in no-till systems may have particularly
signiﬁcant implications for the development of more sustainable agriculture practices.

While this dissertation addresses the role of diversity in row crop
agroecosystems typical of the US Midwest, additional studies will be needed to address
whether other types of agricultural systems may also be amenable to management
through manipulations of crop diversity. There are many examples of diversiﬁed
cropping systems in tropical agroecosystems (Vandermeer 1989, Vanderrneer et al.
1998, Altieri 1999); however, there are far fewer examples of manipulations of diversity
in temperate agricultural systems that do not involve corn, soybean, and wheat (but see
Zhu et a1. 2000).

Future studies should also address the role of weed diversity on crop-weed
competition. For instance, do diverse weed communities compete more or less with
crops? Ecological theory would suggest that a diverse weed community would utilize

more available niches and ﬁrlly exploit available resources and so might preempt

132

resources that would otherwise be available to crops (Tilman 1999). Diverse weed
communities might also be more likely to contain species that are strong competitors
with crops (Tilman 1999, Davis et al. 2005). Conversely, however, diverse weed
communities may be able to suppress dominance among the weed species and so reduce
the competitive impact on the crop. In addition, diverse weed communities may also be
more efﬁcient at sequestering nutrients early and late in the growing season by serving
as ‘natural cover crops’ that prevent soil nutrients from being leached from the system.
Particular weed assemblages might also promote beneﬁcial insects or microbial
associations that would reduce herbivory, disease transmission, or suppress populations
of soil-bome pests (e. g. nematodes). The potentially beneﬁcial effects of weeds in
cropping systems have received relatively little attention (but see Schellhom and Sork
1997, Norris and Kogan 2000), and more work in this area could prove enlightening.
The density at which weed diversity affects crop yields should also be
addressed. There is some indication that at high weed density, diverse weed
communities may have fewer effects on crop yields than less diverse weed communities
at the same density (Gross and Smith, 2003). Future research will be needed to
investigate whether or not this is a general result and the mechanism underlying this

observation.

133

LITERATURE CITED

Altieri, M.A. 1999. The ecological role of biodiversity in agroecosystems. Agriculture
Ecosystems and Environment 74: 19-31.

Andow, D.A. Vegetational diversity and arthropod population response. Annual Review
of Entomology 36, 561-586 (1991).

Cardinale, B.J., Palmer, M.A. 2002. Disturbance moderates biodiversity-ecosystem
function relationships: Experimental evidence from caddisﬂies in stream
mesocosms. Ecology 83: 1915-1927.

Davis, A.S., Renner, K.A., Gross, KL. 2005. Weed seedbank and community shifts in a
long-term cropping systems experiment. Weed Science 53: 296-306.

Donald, W.W., Kitchen, N.R., Sudduth, K.A. 2001. Between-row mowing plus banded
herbicide to control annual weeds and reduce herbicide use in no-till soybean
(Glycine max) and corn (Zea mays). Weed Technology 15: 576-584.

Drinkwater, L.E., Wagoner, P., Sarrantonio, M. 1998. Legume-based cropping systems
have reduced carbon and nitrogen losses. Nature 396: 262-265.

Gross, K.L., Smith, R.G. 2003. Determinants and beneﬁts of weed diversity in row-
cropping systems. Abstracts, Ecological Society of America Annual Meeting,
Savannah, GA.

Huston, M. 1979. General hypothesis of species diversity. American Naturalist 113: 81-
101.

Menalled, F .D., Gross, K.L., Hammond, M. 2001. Weed aboveground and seedbank
community responses to agricultural management systems. Ecological
Applications 1 1: 1586-1601.

Nonis, R.F., Kogan, M. 2000. Interactions between weeds, arthropod pests, and their
natural enemies in managed ecosystems. Weed Science: 94-158.

Reganold, J .P., Glover, J.D., Andrews, P.K., Hinman, HR. 2001. Sustainability of three
apple production systems. Nature 410: 926-930.

Schellhom, N.A., Sork, V.L. 1997. The impact of weed diversity on insect population
dynamics and crop yield in collards, Brassica oleraceae (Brassicaceae).
0ecologia 111: 233-240.

Tilman, D. 1999. The ecological consequences of changes in biodiversity: A search for

134

general principles. Ecology 80: 1455-1474.

Tonhasca, A., Byme, D.N. 1994. The effects of crop diversiﬁcation on herbivorous
insects-a metaanalysis approach. Ecological Entomology 19: 239-244.

Vandermeer, J. 1989. The ecology of intercropping. Cambridge University Press,
Cambridge.

Vandermeer, J ., van Noordwijk, M., Anderson, J ., Ong, C., Perfecto, I. 1998. Global
change and multi-species agroecosystems: Concepts and issues. Agriculture
Ecosystems and Environment 67: 1-22.

Westerman, P.R., Liebman, M., Menalled, F .D., Heggenstaller, A.H., Hartzler, R.G.,
Dixon, PM. 2005. Are many little hammers effective? Velvetleaf (Abutilon
theophrasti) population dynamics in two- and four-year crop rotation systems.
Weed Science 53: 382-392.

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.

Zhu, Y.Y., Chen, H.R., Fan, J.H., Wang, Y.Y., Li, Y., Chen, J .B., Fan, J .X., Yang, S.S.,

Hu, L.P., Leung, H., Mew, T.W., Teng, P.S., Wang, Z.H., Mundt, CC. 2000.
Genetic diversity and disease control in rice. Nature 406: 718-722.

135

APPENDIX A

CHAPTER 4 APPENDIX

136

Appendix A4. 1: Indicator values (IV) of species found in spring (S) and fall-tilled (F)
fallow treatments from 2002-2004. Nativity indicates whether species were introduced

(1) or native UNI) to the USA. Life cycle indicates whether species are annual (A),

biennial (B), or perennial (P). Species in bold are signiﬁcant at P < 0.05 %.

 

 

Species Nativity Life cycle, Disturba IV P
(USA) Growth form nce

Abutilon theophrasti I A, forb S 90.5 0.001
Amaranthus retroﬂexus I A, forb S 41.7 0.034
Ambrosia artemisiifolia I A, forb F 65.4 0.007
Arabidopsis thaliana I A, forb F 89.5 0.001
Aster sp. U A, forb F 16.7 0.491
Barbarea vulgaris I B, forb F 16.7 0.495
Capsella bursa-pastoris I A, forb F 25 0.209
C erastium fontanum I B/P, forb F 16.7 0.483
Chenopodium album U A, forb S 61.2 0.002
Conyza canadensis N A/B, forb F 75 0.001
Daucus carota I B, forb F 25 0.223
Digitaria ischaemum I A, C4 grass S 56.2 0.076
Digitaria sanguinalis N A, C4 grass S 85.9 0.001
Echinochloa crus-galli I A, C4 grass S 8.3 l
Elymus repens I P, C3 grass F 59.1 0.015
Eragrostis cilianensis I A, C4 grass S 41.7 0.034
Erigeron sp. N A/B, forb F 16.7 0.48
Lepidium virginicum N A/B/P, forb F 16.7 0.488
Mollugo verticillata N A, forb S 15.3 0.47
Oxalis stricta N P, forb F 56.8 0.053
Panicum capillare N A, C4 grass F 8.3 1
Panicum dichotomiﬂarum N A, C4 grass S 67.8 0.018
Plantago lanceolata I A/B/P, forb F 33.3 0.096
Plantago major N P, forb F 41.7 0.023
Poa sp. I P, C3 F 41.7 0.029
Polygonum aviculare I A/P, forb F 75 0.002
Polygonum convolvulus I A, forb F 8.3 1
Polygonum persicaria U A/P, forb F 12.2 0.71
Portulaca oleracea N A, forb F 8.3 1
Potentilla norvegica N A/B/P, forb F 25 0.223
Rumex obtusifolius I P, forb F 32.1 0.534
Setariafaberi I A, C4 grass S 75 0.001
Setaria pumila I A, C4 grass S 54.7 0.077
Setaria viridis l A, C4 grass S 66 0.006

 

137

Appendix A4.] (cont’d).

 

 

Species Nativity Life cycle, Disturba IV P
(USA) Growth form nce
Silene latifolia I B/P, forb F 8.3 l
Solanum ptychanthum N A, forb S 8.3 1
Solidago canadensis N P, forb F 33.3 0.105
Stellaria media I A/P, forb F 25 0.225
T araxacum oﬂicinale U P, forb F 81.6 0.001
T rifolium sp. I B/ P, forb F 90.9 0.001
Veronica sp. U A, forb F 15.3 0.45

 

138

   

11111111111111111111111