AN ANALYSIS OF BIOENERGY CROPPING SYSTEMS
IN THE NORTH-CENTRAL UNITED STATES
By
Katherine Elizabeth Hadley

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Crop and Soil Sciences - Master of Science
2014

ABSTRACT
AN ANALYSIS OF BIOENERGY CROPPING SYSTEMS
IN THE NORTH-CENTRAL UNITED STATES
By
Katherine Elizabeth Hadley
Recent U.S. Energy Policy such as the Renewable Fuels Standard has been created to reduce
our dependence on fossil fuels and mitigate global climate change. The GLBRC Bioenergy Cropping
Experiment contains annual cropping systems and perennial cropping systems. Field inputs, gas
emissions, and yields have been tracked since establishment in 2008. Field data was used to
perform a life cycle assessment on these cropping systems with a focus on the global warming
potential (GWP). The results of the LCA showed that the continuous corn cropping systems had the
highest GWP per hectare per year at both locations. The perennial systems generally had a negative
GWP per hectare per year due to limited inputs and gas emissions. Material inputs that contributed
most to GWP were synthetic nitrogen fertilizers. Nitrous oxide emissions had the greatest effect on
the GWP of each system of all of the gas emissions considered. The second chapter of this thesis
research focused on a successful and cost effective establishment method for switchgrass.
Switchgrass is an herbaceous perennial grass capable of producing large amounts of biomass.
Biomass can be converted to cellulosic ethanol. Stand establishment to maximum biomass potential
required several years. A field experiment was conducted to identify if a double crop management
system was possible for switchgrass establishment. The results of the experiment showed that
August seeded switchgrass could not be successfully established following a wheat crop. June
seeded switchgrass after a rye crop and fallow (control treatment) were successfully established.
No significant differences in stand establishment or yields were seen between these two
treatments.

Copyright by
KATHERINE ELIZABETH HADLEY
2014

ACKNOWLEDGEMENTS

Special thanks to Dr. Kurt Thelen for providing me with a great learning experience and
the framework and guidance for this project. I was fortunate enough to have Dr. Karen Renner
and Dr. Steve Hamilton serve as my committee members. Thank you for your time and for your
input and suggestions to make this thesis better.
Thank you to all of the wonderful MSU staff and students who I got to meet along my
journey - Pavani Tumbalam, Andrew Adkins, Randy Laurenz, Casey Reagan, Bill Widdicombe,
Lori Williams, Todd Martin, and Sichao Wang.
A big shout out to Kaite Hard, my partner in crime and other half of the ‘Kates’! I
couldn’t have gotten through grad school without you!
I would also like to thank my parents, friends, and Sigma Alpha sisters who believed in
me and who have supported me throughout my life and this project.

iv

TABLE OF CONTENTS

LIST OF TABLES .............................................................................................................................. viii
LIST OF FIGURES ............................................................................................................................... x
CHAPTER 1

LIFE-CYCLE ASSESSMENT OF THE GREAT LAKES BIOENERGY CENTER’S
BIOENERGY CROPPING SYSTEM EXPERIMENT ........................................................1
ABSTRACT ........................................................................................................................................ 1
INTRODUCTION ............................................................................................................................... 2
Renewable Fuels ................................................................................................................. 2
Life Cycle Assessment ......................................................................................................... 5
GLBRC Biofuels Cropping Experiment................................................................................. 8
Objective ............................................................................................................................. 9
MATERIALS AND METHODS ............................................................................................................ 9
Field Experiment ................................................................................................................. 9
Main Plot Treatments ........................................................................................... 12
Annual Cropping System Treatments ............................................................. 12
Perennial Cropping System Treatments ......................................................... 13
Treatment Descriptions ........................................................................................ 14
G1 - Continuous Corn ...................................................................................... 14
G2 - Continuous Corn + Cover Crops .............................................................. 14
G2 - Corn - Soybean - Canola rotation / G3 - Soybean - Canola - Corn rotation
/ G4 - Canola - Corn - Soybean rotation.................................................... 14
G3 - Corn-Soybean Rotation + Cover Crops / G4 - Soybean-Corn Rotation +
Cover Crops ............................................................................................... 15
G5 - Switchgrass .............................................................................................. 15
G6 - Miscanthus .............................................................................................. 16
G7 - Native Grass mix ...................................................................................... 16
G8 - Hybrid Poplar ........................................................................................... 16
G9 - Old Field................................................................................................... 17
G10 - Native Prairie ......................................................................................... 17
Goal and Scope Definition ................................................................................................ 18
Goal ....................................................................................................................... 18
Temporal and Geographic Scopes ........................................................................ 18
System Boundary .................................................................................................. 19
Co-Products and Allocation .................................................................................. 20
Life Cycle Inventory (LCI)................................................................................................... 20
General Management Practices ........................................................................... 20
LCI Calculation ....................................................................................................... 23
Treatments G1, G5, G6, G7, G9, G10 .............................................................. 23
Rotational Treatments G2, G3, G4.................................................................. 24
v

Greenhouse Gas Fluxes ......................................................................................... 25
Estimating N2O emissions ............................................................................... 26
Gas Flux and Models ....................................................................................... 27
GaBi 6.0 Professional + Extension Database......................................................... 28
Life-Cycle Impact Assessment (LCIA) ................................................................................ 39
TRACI Impact Assessment Methodology ........................................................ 40
Global Warming Potential Impact Characterization ....................................... 42
RESULTS......................................................................................................................................... 43
Life Cycle Interpretation ................................................................................................... 43
Global Warming Potential..................................................................................... 43
Nitrous Oxide Emissions and Global Warming Potential................................ 43
Material Inputs and Global Warming Potential .............................................. 48
Global Warming Potential on a Land Area Basis ............................................ 50
Yield....................................................................................................................... 51
DISCUSSION................................................................................................................................... 55
Nitrous Oxide Emissions ................................................................................................... 55
Fertilizer Sources................................................................................................... 56
Corn Stover ........................................................................................................... 57
Perennial Crops ..................................................................................................... 58
Miscanthus ...................................................................................................... 58
Old Field Treatment ........................................................................................ 59
Land Use Changes ................................................................................................. 60
Model Assumption Effects and Considerations .................................................... 61
Miscanthus Rhizomes ..................................................................................... 62
Crop Residues as a Co-Product ....................................................................... 63
CONCLUSION ................................................................................................................................. 63
APPENDIX ...................................................................................................................................... 66
REFERENCES .................................................................................................................................. 92
CHAPTER 2

MAXIMIZING LAND USE DURING SWITCHGRASS ESTABLISHMENT IN THE
NORTH CENTRAL UNITED STATES..........................................................................95
INTRODUCTION ............................................................................................................................. 96
MATERIALS AND METHODS .......................................................................................................... 98
Treatments, Experimental Design, and Site Description .................................................. 98
Small Grains Management................................................................................................ 99
Switchgrass Management ............................................................................................... 101
Forage Sampling.............................................................................................................. 102
Soil Moisture ................................................................................................................... 103
Switchgrass Stand Frequency ......................................................................................... 103
Theoretical Ethanol Yield ................................................................................................ 104
Statistical Analysis ........................................................................................................... 105
RESULTS AND DISCUSSION ......................................................................................................... 106
Seeding Year Switchgrass................................................................................................ 106
vi

Surface Soil Moisture ...................................................................................................... 108
Subsurface Soil Moisture ................................................................................................ 109
Established Switchgrass: Frequency ............................................................................... 110
Established Switchgrass: Production .............................................................................. 110
Double-Cropping and Dual-Use Cropping Systems ........................................................ 112
Biomass Quality and Theoretical Ethanol Yield .............................................................. 113
CONCLUSION ............................................................................................................................... 116
ACKNOWLEDGEMENTS ............................................................................................................... 117
CHAPTER 2 NOTE ........................................................................................................................ 117
APPENDIX .................................................................................................................................... 120
REFERENCES ................................................................................................................................ 122

vii

LIST OF TABLES

Table 1. Annual cropping system crops by year. Variety information is included in the
appendix........................................................................................................................................ 12
Table 2. Perennial cropping system crops by year. ..................................................................... 13
Table 3. An overview of the management practices and inputs used in each cropping
system by location. ....................................................................................................................... 22
Table 4. Results of the Student’s t-test for testing the difference in nitrous oxide emissions
between locations for each cropping system (α=0.10). ............................................................... 45
Table 5. Species planted in the perennial cropping systems. ...................................................... 71
Table 6. Annual crop variety by year at KBS. ............................................................................... 72
Table 7. Annual crop variety by year at ARL. ............................................................................... 73
Table 8. Herbicides used at ARL. Chemical and trade names are listed. ..................................... 74
Table 9. Herbicides used at KBS. Chemical and trade names are listed. ..................................... 75
Table 10. Material inputs at KBS per reference flow of 1 ha year-1............................................. 76
Table 11. Non-material inputs at KBS per reference flow of 1 ha year-1. .................................... 77
Table 12. KBS: Yield refers to grain yield at standard moisture content for corn, soybeans,
canola; dry biomass yield of perennial crops. .............................................................................. 78
Table 13. KBS: Carbon dioxide accumulation in the soil for each cropping system. ................... 79
Table 14. KBS: Nitrous oxide emissions from each cropping system. ......................................... 79
Table 15. Farm equipment used in field experiment. ................................................................. 80
Table 16. Material inputs at ARL per reference flow of 1 ha year-1. ........................................... 82
Table 17. Non-material inputs at ARL per reference flow of 1 ha year-1..................................... 84
Table 18. ARL: Yield refers to grain yield at standard moisture content for corn, soybeans,
canola; dry biomass yield of perennial crops. .............................................................................. 85
Table 19. ARL: Nitrous oxide emissions from each cropping system. ......................................... 86

viii

Table 20. Site characteristics from seven environments in southern Wisconsin and lower
Michigan........................................................................................................................................ 99
Table 21. License agreement for article reprint. ....................................................................... 120

ix

LIST OF FIGURES

Figure 1. A visual depiction of life cycle assessment. .................................................................... 6
Figure 2. Arlighton Argicultural Research Station plot map. ....................................................... 10
Figure 3. Kellogg Biological Station plot map. ............................................................................. 11
Figure 4. Cropping system model of treatment G1 Continuous at ARL. ..................................... 29
Figure 5. Cropping system model of corn in rotation (corn-rot) at ARL. ..................................... 30
Figure 6. Cropping system model of soybeans in rotation (soybean-rot) at ARL. ....................... 30
Figure 7. Cropping system model of canola in rotation (canola-rot) at ARL. .............................. 31
Figure 8. Cropping system model of normalized corn-soybean-canola rotation (rotational
system) at ARL. .............................................................................................................................. 32
Figure 9. Cropping system model of G5 Switchgrass at ARL. ...................................................... 32
Figure 10. Cropping system model of G6 Miscanthus at ARL. ..................................................... 33
Figure 11. Cropping system model of G7 Native Grass mix at ARL. ............................................ 33
Figure 12. Cropping system model of G9 Old Field at ARL. ......................................................... 34
Figure 13. Cropping system model of G10 Native Prairie mix at ARL.......................................... 34
Figure 14. Cropping system model of treatment G1 Continuous at KBS. ................................... 35
Figure 15. Cropping system model of corn in rotation (corn-rot) at KBS. ................................... 35
Figure 16. Cropping system model of soybeans in rotation (soybean-rot) at KBS. ..................... 36
Figure 17. Cropping system model of canola in rotation (canola-rot) at KBS. ............................ 36
Figure 18. Cropping system model of normalized corn-soybean-canola rotation (rotational
system) at KBS............................................................................................................................... 37
Figure 19. Cropping system model of G5 Switchgrass at KBS. .................................................... 37
Figure 20. Cropping system model of G6 Miscanthus at KBS. ..................................................... 38
Figure 21. Cropping system model of G7 Native Grass mix at KBS. ............................................ 38

x

Figure 22. Cropping system model of G9 Old Field at KBS. ......................................................... 39
Figure 23. Cropping system model of G10 Native Prairie mix at KBS. ......................................... 39
Figure 24. An overview of TRACI 2.0 midpoint impact categories and possible endpoints. ....... 41
Figure 25. Global warming potential pathway starting at the life cycle inventory phase. ......... 43
Figure 26. Average annual nitrous oxide emissions of each cropping system for the study
years 2008-2012. .......................................................................................................................... 44
Figure 27. Nitrous oxide emissions by year for the perennial cropping systems at each
location. ........................................................................................................................................ 47
Figure 28. Nitrous oxide emissions by year for the annual cropping systems at each
location. ........................................................................................................................................ 48
Figure 29. Material input contribution to the global warming potential of each system at
Arlington Agricultural Research Station (ARL). ............................................................................. 49
Figure 30. Material input contribution to the global warming potential of each system at
Kellogg Biological Station (KBS). ................................................................................................... 49
Figure 31. Average global warming potential of each cropping system on a land area basis
per year for the study years 2008-2012 (GWP kg CO2-equiv ha-1 yr-1)......................................... 51
Figure 32. Annual grain yield of annual systems and dry biomass yield of perennial
systems. ........................................................................................................................................ 52
Figure 33. Average estimated energy (MJ) produced on one hectare of each cropping
system per year for study years 2008-2012. ................................................................................ 53
Figure 34. Net global warming potential mitigated (blue bar) by the biofuel crop when
compared to equal amounts of fossil fuel energy produced by each system.............................. 54
Figure 35. Net global warming potential (kg CO2-eq ha-1 yr-1) verses the energy produced,
in MJ, by each cropping system. ................................................................................................... 55
Figure 36. Annual temperature and precipitation for Arlington Agricultural Research
Station (ARL) and Kellogg Biological Station (KBS). ...................................................................... 66
Figure 37. Yearly air temperature and precipitation data from the Hickory Corners weather
station recorded by Michigan State University Enviro-weather service. ..................................... 67
Figure 38. Yearly air temperature and precipitation data from the Arlington University
Farm weather station recorded by the National Weather Service. ............................................. 69

xi

Figure 39. Seeding year switchgrass yields at each of the seven environments....................... 107
Figure 40. Volumetric water content Θ during the 2010 and 2011 growing seasons............... 109
Figure 41. Second year switchgrass yields at each of the seven environments........................ 111
Figure 42. Biomass quality for ethanol production at seven environments across four reps
and two previous crop (i.e., rye or fallow) treatments. ............................................................. 115
Figure 43. Potential ethanol yield at seven environments across four reps and two
previous crop (i.e., rye or fallow) treatments............................................................................. 116

xii

CHAPTER 1
LIFE-CYCLE ASSESSMENT OF THE GREAT LAKES
BIOENERGY CENTER’S BIOENERGY CROPPING SYSTEM EXPERIMENT

ABSTRACT
Recent legislation such as the Renewable Fuels has been created to reduce our
dependence on fossil fuels and mitigate global climate change. The GLBRC Bioenergy Cropping
Experiment, established in 2008 at UW-Madison and MSU, contains annual cropping systems
(corn, soybeans, canola) and perennial cropping systems (switchgrass, miscanthus, a native
grass mix, old field, poplar, a prairie grass mix). Field inputs, gas emissions, and yields have been
tracked since establishment. This field data was used to perform a life cycle assessment on
these cropping systems with a focus on the global warming potential (GWP). The results of the
LCA showed that the continuous corn cropping systems had the highest GWP per hectare per
year at both study locations as expected. The net GWP of a cropping system was heavily
influenced by the nitrous oxide emissions for that system. A statistical analysis revealed no
significant differences between locations for each cropping system (Student’s t-test) or
between cropping systems at each location (ANOVA) likely due to the large variations in annual
nitrous oxide emissions. The perennial systems generally had a negative GWP per hectare per
year due to limited inputs and gas emissions from those systems. The material inputs that
contributed most to GWP were synthetic nitrogen fertilizers. Nitrous oxide emissions had the
greatest effect on the GWP of each cropping system of all of the gas emissions considered.

1

INTRODUCTION
Renewable Fuels
The Energy Independence and Security Act (EISA) of 2007 was an act created and
passed by the US government in order to increase energy security and to promote energy
independence. Part of the EISA included revisions and updates to the Renewable Fuels
Standard (RFS) program. The RFS already included framework for developing biofuels as an
alternative fuel source to reduce greenhouse gas (GHG) emissions and mitigate global climate
change (Menten et al, 2013). The newly revised program, referred to as RFS2, redefined the
criteria for renewable fuels and set volume requirements for use as transportation fuels each
year. The RFS2 requires the amount of renewable fuels produced in the US to increase each
year leading up to 36 billion gallons per year (BGY) in 2022 (US Department of Energy, 2011). In
addition to the total volume requirements, RFS2 outlines specific fuel source requirements: 16
BGY of cellulosic biofuels, 14BGY of advanced fuels, 1 BGY of biomass-based biodiesel, and 15
BGY of conventional biofuels (US Department of Energy 2011). The new definition of renewable
fuels also includes criteria for greenhouse gas emission reduction thresholds.
Starch-based ethanol and biodiesel are conventional biofuels, also called first
generation biofuels, and are the primary liquid transportation biofuels available in the US.
Currently, over 95% of the ethanol produced in the US is from corn grain (Zea mays) (Drapcho,
2008). Biodiesel production has traditionally used soybeans (Glycine max) as the main
feedstock source. In 2007, 83% of biodiesel was soybean biodiesel (US Department of Energy,
2011). Other oil seed crops, such as canola (Brassica napus) or sunflowers (Helianthus annuus),
are also used as biodiesel feedstocks, but to a much lesser extent.
2

From a sustainability standpoint, first generation biofuels present two major challenges:
large agronomic input requirements during the production phase and land use concerns.
Conventional commodity cropping systems use many agronomic inputs including high rates of
commercial fertilizers. The production and use of fertilizers significantly contribute to the
environmental impacts of a cropping system because of the raw material extraction and energy
use required for their production (Skowroñska, 2014). Future cropping systems will have to
become more efficient in their use of agronomic inputs to be sustainable in the long term.
Sustainable production is necessary because biofuels produced from conventional sources are
required to reduce GHG emissions by 20% as compared to the lifecycle GHG emissions of the
2005 baseline average gasoline or diesel fuel that it replaces to be considered renewable fuels
under the RFS2 definition (Congressional Research Service, 2013). Using arable land for biofuel
feedstock production instead of food production has been a source of controversy for
policymakers. In 2010, 38% of corn grain produced in the US was used for ethanol production,
totaling 13.2 billion gallons of ethanol and 39 million metric tons of dry distiller grains (US
Department of Energy, 2011; ethanolrfa.org, 2014). Dry distiller grains (DDG) are an important
co-product of ethanol production and are used as a high quality livestock feed. According to the
Renewable Fuels Association, nearly one-third of the corn grain used for ethanol production is
not removed from the food market. Demand has grown for ethanol in the US due to increased
gasoline blend requirements (E-15) and the introduction of Fuel Flex vehicles (E-85). As the corn
grain ethanol market continues to expand, food vs. fuel becomes a concern due to biofuels
feedstock production taking acres away from food production. With the world’s population

3

expected to increase to 9+ billion people by 2050, food production will have to increase 50-80%
during the same time period (Brentrup, 2008).
Cellulosic derived ethanol, a second generation biofuel, has been proposed as a viable
alternative to using traditional food crops as biofuel feedstocks. Cellulosic ethanol is produced
by breaking down lignocellulos in plant biomass to simple sugars, mainly glucose and xylose.
The simple sugars are fermented into ethanol. Feedstocks for cellulosic ethanol have been in
development since 1976 when the US Department initially funded research programs to
identify high-yielding herbaceous biomass crops (Wright, 2007). In 1991, switchgrass (Panicum
virgatum) was identified as a ‘model’ bioenergy crop (US Department of Energy, 2011). Since
the early 1990’s various other grass and woody species have been identified as potential
bioenergy feedstocks as well.
The purpose of cellulosic derived ethanol is that it has the potential to alleviate the
sustainability concerns caused by first generation biofuels. Dedicated bioenergy crops are not
food crops. To further reduce food vs. fuel concerns, cellulosic feedstocks have been bred to be
grown on marginal lands where sustainable food production is unlikely or impossible. Many
cellulosic bioenergy crops do not require large amounts of agronomic inputs, such as fertilizers
or herbicides, although these requirements vary by crop. High-yielding, low input biomass
systems can potentially make cellulosic ethanol able to meet the RFS2 GHG emission reduction
threshold of a 60% reduction as compared to the lifecycle GHG emissions of the 2005 baseline
average gasoline or diesel fuel that it replaces (Congressional Research Service, 2013).
Any biofuel that is not corn starch-based qualifies as an advanced biofuel under the
RFS2. Advanced biofuels can come from grain (other than corn), cellulosic sources (including

4

corn stover), and biomass based biodiesel (Congressional Research Service, 2013). The inclusion
of advance biofuels in the RFS2 requirements bridge the gap between first generation biofuels
and second generation biofuels and allow for a transition period while technologies and
infrastructure are developed for cellulosic biofuels. Advanced biofuels also allow for a
mandated GHG emission reduction threshold during the transition period - a 50% reduction as
compared to the lifecycle GHG emissions of the 2005 baseline
average gasoline or diesel fuel that it replaces (Congressional Research Service, 2013).

Life Cycle Assessment
Life cycle assessment (LCA) is a decision making tool used to assess environmental
impacts of a product or system. Historically, LCA has been used in the packaging and
manufacturing industries where material inputs are readily calculated and factory and process
emissions can be easily identified and quantified for each life cycle stage. Due to the
overwhelming success of LCA studies in these industries, LCA work has expanded into other
applications such as agricultural systems. The first biofuel LCA was published by Kaltschmitt et
al in 1997 and focused on first generation biofuels. After that initial study, LCA had become an
accepted and highly used approach to study environmental impacts, including GHG emissions
of biofuels (Menten et al, 2013).
During a life cycle assessment, the entire life cycle of a system is broken down into all of
its life stages starting with the ‘cradle’, extraction of resources, to the ‘grave’, end of life. Each
life cycle stage is examined and weak stages can be identified. The principles and framework for
life cycle assessment are defined in ISO 14040:2006 with additional requirements and

5

guidelines defined in ISO 14044:2006. A complete LCA is composed of four main phases: (1)
goal and scope definition, (2) life cycle inventory, (3) life cycle impact assessment, and (4) life
cycle interpretation. LCA phases are interconnected as shown in Figure 1.

Figure 1. A visual depiction of life cycle assessment.

Other than the general guidelines and requirements outlined in ISO standards, there is
no standardized methodology for conducting life cycle assessments. The lack of standardization
in LCA makes it hard to compare results of one study to another. Agriculture LCAs also face a
unique set of challenges that cause a great deal of variation and uncertainty in the results
(Kendall et al, 2013). Obvious differences such as climate, soil type and health, and
management practices vary greatly by region and can have a major impact on the agronomic
inputs and yields of cropping systems. Many biofuel LCAs rely on data from computer simulated
cropping systems such as the CERES-EGC model to estimate crop yields and environmental

6

emissions for their life cycle inventory (LCI) (Bessou, 2012). Very few LCAs review used field
data and/or field measured gas fluxes in their life cycle inventories. The use of experimentally
determined field data can improve the data quality component of the LCI in cropping system
LCAs and tailor results to specific geographic regions and management practices.
The type of LCA and other LCA methodology choices will also impact the results of a
study. There are two types of LCAs, attributional and consequential. LCAs have traditionally
been conducted as attributional LCAs. Usually retrospective, this method simply evaluates the
inputs and outputs of life cycle and identifies areas of greatest potential environmental impact.
Consequential LCAs are usually prospective and consider changes to a life cycle in order to
reduce potential environmental impacts. Consequential LCAs also consider direct and indirect
impacts of changing the life cycle (Kendall et al, 2013). The effects of indirect land use changes
are especially important to biofuel production but are hard to study even with consequential
LCAs. Consequential LCAs require significantly more time, and resources, and are more costly,
yet still can have high levels of variability and uncertainty due to direct and indirect system
changes. The LCA conducted in this research will be a retrospective attributional LCA and will
not account for land use changes cause by dedicated biofuel crops replacing food crops on
arable land in the study locations.
Functional unit choice has been shown to significantly impact the results of a biofuel
LCA (Caffrey, 2012). The functional unit is a measure of the intended function for the system.
Biofuel LCAs commonly use one of three functional units: energy (ie. MJ of energy), land area
(ie. hectare of land), or mass (ie. kg of corn grain). Mass based functional units are used when
comparing systems with the exact same product; for example comparing switchgrass harvest

7

strategies would use a function unit of kilograms of dry switchgrass biomass. Energy functional
units are useful when comparing feedstocks that produce different products; for example
comparing corn grain ethanol to soybean biodiesel might use a functional unit of MJ of energy.
Land area functional units are used when the intensity of cropping systems is to be compared
(Skowronska et al, 2014). A land area functional unit will be primarily utilized in the LCA
conducted in this research. Using a land area functional unit also increases the practicality of
the research in terms of the future bioenergy feedstock considerations from the results.

GLBRC Biofuels Cropping Experiment
The Great Lakes Bioenergy Research Center (GLBRC) is one of three bioenergy research
centers created and funded for five years by the US Department of Energy (DOE) in 2007. In
2013, funding was extended for an additional five years. The GLBRC mission statement is as
follows: “to perform the basic research that generates technology to convert cellulosic biomass
to ethanol and other advanced biofuels.” (GLBRC, 2014). The GLBRC is broken down into four
main areas of research Area 1: Plants, Area 2: Deconstruction, Area 3: Conversion, and Area 4:
Sustainability. Each area is further divided into specific areas of interest. Through collaboration
between research areas, large-scale research is made possible. One such large-scale research
effort is the GLBRC Biofuels Cropping System Experiment.
The goal of the GLBRC Biofuels Cropping System Experiment is to model the production
efficiencies of novel bioenergy crop production systems. The experiment includes “assessing
crop yield and quality, microbial-plant interactions, biogeochemical responses, water needs,

8

greenhouse gas fluxes, nutrient leaching to groundwater, biodiversity responses and
socioeconomic impacts” (GLRBC, 2014).

Objective
The objective of this research is to perform an attributional life-cycle assessment to
evaluate the global warming potential of the production phase the main plot treatment
cropping systems located in the GLBRC Biofuels Cropping Experiment. The life-cycle inventory
will primarily consist of the actual field data acquired between experiment establishment in
2008 and the end of the growing season in 2012. Field data includes material inputs and nonmaterial inputs. Experimentally derived nitrous oxide (N2O) gas flux data will also be used in the
life-cycle inventory. Life-cycle assessment software GaBi 6.0 (P.E. International) will be used to
create representative models of each cropping system.

MATERIALS AND METHODS
Field Experiment
The GLBRC Biofuels Cropping System Experiment intensive sites were established in
2008 at the Arlington Agricultural Research Station (ARL) (43°18’10.2”N, 89°20’43.0” W) in
Arlington, Wisconsin with a second location at Kellogg Biological Station (KBS) (42°24’19”N,
85°24’04”W) in Hickory Corners, Michigan. The experimental design is randomized complete
block design with five blocks. The experiment has ten cropping system treatments. Plots are
approximately 40 meters long by 28 meters wide with an area of 0.12 hectares. Field layouts
are shown in Figures 1 and 2. Prior to experiment establishment in 2008, alfalfa was grown on

9

the entire KBS site and on blocks 1, 2, and 3 at ARL. Blocks 4 and 5 at ARL had corn as the
previous crop. Field activities from both locations have been carefully tracked and recorded for
each system since establishment in 2008.

Figure 2. Arlighton Argicultural Research Station plot map.

10

Figure 3. Kellogg Biological Station plot map.

11

Main Plot Treatments
Annual Cropping System Treatments
Table 1.. Annual cropping system crops by year. Variety information is inc
included
luded in the appendix.

12

Perennial Cropping System Treatments
Table 2. Perennial cropping system crops by year.
System

System Name

Group

G5

Switchgrass

Graminoids

G6

Miscanthus

G7

Native Grass Mix

G8

Hybrid Poplar

G10

Prairie Grass Mix

Species (latin)

Common Name / Variety

C4

Panicum virgatum

switchgrass / Cave-in-rock

Graminoids

C4

Miscanthus x giganteus

giant miscanthus

Graminoids

C3

Elymus canadensis

Canada wildrye

C4

Panicum virgatum

switchgrass

Andropogon gerardii

big bluestem

Schizachyrium scoparium

little bluestem

Sorghastrum nutans

Indiangrass

Populus nigra x P.
maximowiczii

Hybrid poplar / NM6

Koeleria cristata

prairie Junegrass

Elymus canadensis

Canada wildrye

Panicum virgatum

switchgrass

Andropogon gerardii

big bluestem

Schizachyrium scoparium

little bluestem

Sorghastrum nutans

Indiangrass

Desmodium canadense

showy ticktrefoil

Lespedeza capitata

roundhead lespedeza

Baptisia leucantha

white false indigo

Rudbeckia hirta

blackeyed Susan

Anemone canadensis

Canadian anemone

Asclepias tuberose

butterfly milkweed

Silphium perfoliatum

cup plant

Monarda fistulosa

wild bergamot

Ratibida pinnata

pinnate prairie coneflower

Solidago rigida

rigid goldenrod

Solidago speciosa

showy goldenrod

Aster novae-angliae

New England aster

Graminoids

C3

C4

Legumes

Early forbs

Mid forbs

Late forbs

13

Treatment Descriptions
G1 - Continuous Corn
Continuous corn is a common system in the Midwestern United States. The purpose of
this treatment is to represent an annual high-input cropping system and act as a basis for
comparison with other systems.

G2 - Continuous Corn + Cover Crops
In the 2012 growing season, the canola phase was eliminated from the experiment
because canola isn’t a widely grown crop the Midwest. The change from a corn-soybean-canola
rotation to a corn-soybean rotation better reflects current cropping practices in the region. To
replace the rotation, an annual conventional corn with cover crops system was added.
Conventional corn with cover crops will provide a comparison for treatment G1. Cover crops
provide cover during winter months reducing erosion and can provide nitrogen credits to the
system if legumes such as red clover are used as cover crops.

G2 - Corn - Soybean - Canola rotation / G3 - Soybean - Canola - Corn rotation / G4 - Canola Corn - Soybean rotation
Treatments G2, G3, and G4 had the same annual corn-soybean-canola rotation systems
in 2008, 2009, 2010, and 2011, but each treatments represents a different entry point of the
rotation. The purpose of this is to have one treatment of corn in rotation, one treatment of
soybeans in rotation and one treatment of canola in rotation each study year. The introduction
of canola into the basic corn-soybean rotation increases the biodiversity of the rotation, and if
winter canola varieties are used can provide vegetative cover during winter months to reduce

14

erosion. Canola biodiesel has a lower gel temperature than soybean biodiesel which can make
canola biodiesel favorable in cold environments.

G3 - Corn-Soybean Rotation + Cover Crops / G4 - Soybean-Corn Rotation + Cover Crops
The canola phase was removed in 2012 and future experiment years to accommodate
the implementation of the cover crop systems. An annual corn-soybean rotation is more
common than corn-soybean-canola rotation in the Midwestern US and therefore the new
treatment is a better representation of cropping systems found in the region. Cover crops were
also included in this treatment for the same reasons listed above.

G5 - Switchgrass
Switchgrass was identified by the US DOE as a ‘model’ bioenergy crop in 1991 and is still
considered a bioenergy feedstock of major importance (US Department of Energy, 2011).
Native to the US, switchgrass is a perennial grass capable of producing up to 10 tons of biomass
per acre in northern climates (MSU Extension, 2007). Switchgrass is direct seeded. Issues can
arise with hard seed, so seeding rates are generally inflated to ensure even stand emergence
(Hedtcke et al, 2014). Stand establishment can take 2-3 years to reach peak biomass yields. A
first year biomass harvest is not recommended and could affect the overwintering ability of the
stand. First year management generally includes weed control (herbicide or mowing) but no
fertilizer application. Management in subsequent years requires only fertilizer application,
specifically nitrogen. Switchgrass can be used for cellulosic ethanol or co-fired with coal.

15

G6 - Miscanthus
Giant miscanthus (Miscanthus x giganteus) is a cross between Miscanthus sinensis and
Miscanthus sacchariflorus and is a prime example of hybrid vigor. Giant miscanthus produces
more biomass than either of its parents and is sterile. Miscanthus must be propagated by
rhizome which reduces concerns of it becoming an invasive species. Miscanthus is a perennial
crop capable of producing large amounts of biomass with yields reaching 10-15 tons per acre in
northern climates (MSU Extension, 2007). Stand establishment may take several years to reach
peak biomass. A first year biomass harvest is not recommended and could affect the
overwintering ability of the stand. First year management generally includes weed control
(herbicide or mowing) but no fertilizer application. Management in subsequent years requires
only fertilizer application, specifically nitrogen. Although miscanthus is a viable feedstock for
cellulosic ethanol, the relatively high lignin content makes it a better direct combustion fuel
source with current conversion technologies.

G7 - Native Grass mix
The Native Grass mix contains five perennial grass species native to the US. See Table 2
for complete species list. The purpose of G7 is to serve as an intermediate between the highly
diverse G10 system and the monoculture G5 switchgrass system. Biomass harvested at the end
of each season could theoretically be used for cellulosic ethanol production.

G8 - Hybrid Poplar
Hybrid poplar is a fast growing tree species and is the only woody species included in
this cropping experiment. Hybrid poplar is also the only crop in this experiment that is not

16

harvested each year. Hybrid poplar is vegetatively propagated by cuttings called whips.
Predation by animals and weed competition can be an issue during the initial establishment
phase and must be controlled to ensure even stand establishment. Hybrid poplar can be used
as a cellulosic feedstock however the moderately high lignin content makes conversion under
current technologies not ideal. Genetically modified hybrid poplar varieties with reduced lignin
contents are currently in development.

G9 - Old Field
As opposed to the high-input continuous corn G1 treatment, the old-field treatment is
the low-input treatment in this experiment used for comparison with other systems. The oldfield treatment was tilled at establishment (as was every treatment) in 2008 but nothing was
planted. Whatever seeds were present in the soil seed bank were able to germinate and grow
unregulated. The Old-field treatment is considered a perennial treatment because the seed
bank and essentially the weeds in the plots produce seed to grow the next years ‘crop’. This
treatment received annual nitrogen application so it cannot be considered an untreated
control. Biomass harvested at the end of each season could theoretically be used for cellulosic
ethanol production.

G10 - Native Prairie
The Native Prairie treatment is the high-diversity, low input (HDLI) treatment in this
experiment. This treatment contains a perennial mixture of six grasses, three legumes, three
early forbs, three mid forbs and three late forbs. See Table 2 for complete species list. All of the
species in this treatment are native to the US and would have been found naturally across the

17

Midwestern prairies before production agriculture. This treatment did not receive
supplemental nitrogen fertilizer. Biomass harvested at the end of each season could
theoretically be used for cellulosic ethanol production.

Goal and Scope Definition
Goal
The goal of this research was to conduct an attributional life-cycle assessment (LCA) to
identify and compare the global warming potential of the main plot treatment in each cropping
systems in the GLBRC Biofuels Cropping Experiment on a land area basis. The functional unit is
one hectare of arable land per year. Systems with the lowest environmental impacts,
specifically, the lowest global warming potential, will be highlighted for future research.

Temporal and Geographic Scopes
The time frame for this research is from experiment establishment in the spring of 2008
through harvest in 2012.
Experimental locations were located in the north central US, therefore management
practices and the grain and biomass yields are reflective of the climate and soil conditions of
the region. During the study years, the average annual temperature was 6.8°C and average
annual precipitation was 913 mm at the ARL location (National Weather Service, 2014). The
average annual temperature was 9.2°C and average annual precipitation was 878 mm at the
KBS location (MSU Enviro-weather, 2014). Plano silt-loam (Family: Fine-silty, Mixed,
Superactive, Mesic Typic Argiudolls) is main soil series present at the ARL experiment site. This
soil was formed under prairie vegetation in loess deposits over calcareous glacial till and
18

contained 3.8% soil organic matter (SOM). The drainage class of this soil is well drained.
Kalamazoo loam (Family: Fine-Loamy, Mixed, Semiactive, Mesic Typic Hapludalfs) is the main
soil series present at the KBS experiment site. This soil was formed under forest vegetation in
loamy outwash overlaying sand and gravel and contained 2.4% SOM. The drainage class of this
soil is well drained. Soil information is from an unpublished manuscript from Gregg Sanford at
the University of Wisconsin – Madison.
Hybrid poplar, treatment G8, was intentionally left out of this study. Hybrid poplar was
established when the experiment began 2008 but was not completely harvested within the
timeframe of this study. Hybrid poplar was harvested in late 2012 at ARL and early 2013 at KBS.

System Boundary
Because the goal of this study is specific to the cropping system phase of biofuel
feedstock production, the system boundary will be cradle-to-field gate. Using cradle as the
starting point of this study ensured that the extraction and production of agronomic inputs was
considered. However, the labor in man hours and all other factors pertaining to the farm
workers, including but not limited to transportation to work, calories consumed, human health
index, were not considered in this study. Field gate is the point at which crops were harvested
from the field but have not left the field or been transported away from the field.
Transportation to storage facilities on the farm, grain elevators, biomass processing facilities, or
bio-refineries is outside of the system boundary for this study because it is dependent on the
type of feedstock harvested (grain or biomass). Differences in feedstock transportation should
be considered in life-cycle assessments where conversion is the end point in the system
boundary (ie. not this study).
19

Co-Products and Allocation
Identification of co-product allocation procedures in LCAs is required by the ISO
standards. Due to the complex nature of agricultural systems, allocation procedures are rarely
consistent between studies and are at the discretion of the researcher. In this study, the only
co-product was the crop residue returned to the soil after grain and biomass harvest. Crop
residue from the previous year was considered part of the soil due to cropping system
modeling limitations imposed by the LCA software. Crop nutrition requirements were
calculated based on soil sampling each year and therefore included nutrients returned to the
soil from the previous crops residue. A similar approach was used by Bessou et al. 2012 for
modeling sugar beet cropping systems.

Life Cycle Inventory (LCI)
The life cycle inventory (LCI) for the LCA used as much experimentally derived data as
possible. Experimentally derived data comes from the field data and observed gas fluxes from
the GLBRC Biofuel Cropping System Experiment. Field data including but not limited to: material
inputs, field operations, and yield data from each location, was carefully tracked and recorded
since experiment establishment. With authorization, field data logs can be accessed at
data.sustainability.glbrc.org under the heading ‘GLBRC Biofuel Cropping System Experiment’.

General Management Practices
Crop management decisions were based on best management practices and Extension
recommendations for each location and were dependent on local climate and soil properties.
Nitrogen rates for corn systems were calculated using the Maximum Return to Nitrogen tool
20

and depended on soil type, previous crop, yield expectation, nitrogen cost, and corn price.
Annual crop variety information can be found in Table 6 of the Appendix. Soil tests were
conducted at the beginning of each growing season and together with rotation sequence (in the
annual cropping systems) were used to determine fertilizer requirements each year. Soil pH
was adjusted in all plots with dolomitic lime prior to experiment establishment at ARL. Based on
soil test results, pH adjustment was not required at KBS.
Tillage operations were performed only in 2008 at experiment establishment on all
plots. Tillage was more extensive at ARL than KBS. At ARL, a chisel plough was used as primary
tillage. During secondary tillage plots were disked, field cultivated and culti-packed. At KBS,
secondary tillage from a soil finisher was the only tillage operation. No-till planting and
management practices were used in subsequent years for all of the annual treatments.
Production scale equipment was used to harvest all crops. Corn and soybean harvest
occurred after physiological maturity (black layer in corn, 95% of pods reached full color in
soybeans) and when the grain reached optimum moisture content for grain quality and safe
storage. Canola was swathed at 60% seed color change and then combined five to fourteen
days later. Perennial crops were harvested within two weeks of the first killing frost (-3°C) of
the fall. At ARL, perennial crops were windrowed then chopped with a self-propelled forage
harvester. At KBS, perennial crops were direct chopped with a self-propelled forage harvester.
All perennial crops were cut a 15 cm stubble height.
Table 3 presents an overview of the management practices and inputs used in each
cropping system by location. The solid grey color indicates the systems received that particular
input.

21

Table 3. An overview of the management practices and inputs used in each cropping system by location. The solid grey color
indicates the systems received the input indicated in the first column.
System
Crop
Location
Average Seeding
-1
Rate (ha )
Row Spacing
Ammonium
Polyphosphate
(10-34-0)
Monoammonium
Phosphate
(11-52-0)
Custom Starter
(5-14-42)
Urea Ammonium
Nitrate
(28-0-0)
Ammonium
Nitrate
(34-0-0)
Urea
(46-0-0)
Potash
(0-0-60)
Weed
Control

G1 / corn-rot

Soybean-rot

Canola-rot

G5

G6

Corn

Soybean

Canola

Switchgrass

Miscanthus

ARL
KBS
6.46
11.34
kg
kg
19 cm

ARL

ARL

KBS

17,200

16,714

rhizomes

rhizomes

ARL
84,000
sds
76 cm

KBS
70,900
sds

ARL

KBS

432,000
sds

444,800
sds

38 cm

7.6 kg

22

KBS

G7
Native Grass
Mix
ARL
KBS

G9

ARL

8.9 kg

0 kg

Old Field
KBS

G10
Prairie Grass
Mix
ARL
KBS
8.5 kg

LCI Calculation
Treatments G1, G5, G6, G7, G9, G10
Each material input was averaged across the five treatment years to come up with one
representative value to plug into the cropping system models. Field data used Imperial
measurement units so everything was converted to metric units as well (ie. kg, ha). For
example, G1-Continuous Corn at KBS was planted at 28,000 seeds acre-1 in 2008, 28,000 seeds
acre-1 in 2009, 28,000 seeds acre-1 in 2010, 29,300 seeds acre-1 in 2011, and 30,200 seeds acre-1
in 2012. The value for corn seed in the G1-Continuous Corn at KBS model is 20.05 kg.
ሺ28,000 ൅ 28,000 ൅ 28,000 ൅ 29,300 ൅ 30,200ሻ ൊ 5 years ൌ 28,700 seeds acre-1
28,700

acre
seeds
1
ൈ
ൌ 70,919 seeds hectare-1
acre 0.404686 hectare

൬ሺ70,919 ൈ 56

lbs
seeds
1 kg
ሻ ൊ 90,000
൰ൈ
ൌ 20.05 kg hectare-1
bu
bu
2.2 lbs

A complete list of material input values used in the cropping system models is located in
Appendix Table 16 for ARL and Appendix Table 10 for KBS. Estimated diesel use was calculated
for each piece of equipment based on equipment type and horsepower using the MSU
Extension Custom Machine and Work Rate Estimates for 2013. A complete list of equipment
used can be found in Appendix Table ##.
Non-material inputs included all field operations, tillage, planting, fertilizing, weed
control (spraying), and harvesting, and arable land (recall, 1 ha year-1 is the functional unit).
Non-material inputs were averaged across the five treatment years to come up with one
representative value to plug into the cropping system models the same way as the material

23

input values were calculated. A complete list of non-material input values used in the cropping
system models is located in Appendix Table 17 for ARL and Appendix Table 11 for KBS.
Yields were also averaged across the five treatment years. The average corn grain yield
in Mg was calculated for G1-Continuous Corn. Stover was not considered in this study. Mg of
dry matter biomass was used for the perennial treatments. Treatments were primarily studied
and compared based on a land area basis (1 ha year-1), however, treatment yields were
converted to predicted biofuel energy in MJ and compared on a per energy basis. The purpose
of this was to show that functional unit choice would significantly impact the results of the LCA.
To convert yields (Mg) to energy (MJ), equations [1], [2], [3], and [4] in the Appendix were used.
Continuous corn produced the greatest amount of energy (in MJ) at each location and set to a
factor of 1. The other treatment cropping system models were scaled appropriately based on
their estimated energy production in MJ compared to the continuous corn treatment.

Rotational Treatments G2, G3, G4
Rotational treatment material and non-material input model values were calculated in
much of the same way as the other treatment model values. The difference was that the
rotational systems were separated into models for each crop in the rotation. Corn in rotation
(corn-rot) was an average of six treatment years: G2-2008, G2-2011, G2-2012, G3-2010, G32012, and G4-2009. Soybeans in rotation (soybean-rot) was an average of five treatment years:
G2-2009, G3-2008, G3-2011, G4-2010, and G4-2012. Canola in rotation (canola-rot) was an
average of four treatment years: G2-2010, G3-2009, G4-2008, and G4-2011.
A rotational system model was also created: the corn-rot values were multiplied by six
(treatment years), the soybean-rot values were multiplied by five, and the canola-rot values
24

were multiplied by four. Each input value was summed and then divided by fifteen treatment
years.
Material and non-material input values for the corn-rot, soybean-rot, canola-rot, and
the rotational system cropping system models for each location are also located in Appendix
Tables 10, 11, 16, and 17.

Greenhouse Gas Fluxes
Carbon dioxide, nitrous oxide and methane are the major greenhouse gases that
contribute to global warming. Methane emissions from cropping systems have been proven to
have an insignificant affect on global warming potential compared to carbon dioxide and
nitrous oxide (Bessou et al. 2012, Menten et al. 2013, Skowronska et al. 2014). For this reason,
the methane gas fluxes of the cropping systems were not considered in this study.
All of the treatments in this study except for treatment G10 and the soybean phase of
the rotation systems received nitrogen fertilizer applications. Nitrous oxide is naturally
produced in soils during nitrification and de-nitrification (Huang, 2014). However, the most
important nitrogen emission is nitrous oxide (N2O) in terms of global warming potential
because of its radiative effect in the atmosphere. One kg of nitrous oxide has the effect of 298
kg of carbon dioxide on global warming potential. Global warming potential is expressed in
units of carbon dioxide equivalents (CO2-equiv, CO2-e). Studying nitrous oxide gas fluxes will
help researchers evaluate options for emission mitigation through management practice
changes and/or policy intervention. Nitrous oxide emissions from the cropping systems were
experimentally determined by the GLBRC Area 4.3 Biogeochemical Research Group.

25

Estimating N2O emissions
The N2O emission estimation protocol is an excerpt from Oates et al. (unpublished
2004). “While soil temperatures were consistently > 0 C, nitrous oxide (N2O) fluxes were
measured twice monthly, as well as immediately following precipitation events (Robertson et
al., 2000) using vented static chambers (Livingston & Hutchinson, 1994). The chambers had a
28.45 cm diameter and 18 cm headspace height and were inserted 5 cm below the soil surface.
Chamber lids were modified with a septum for gas extraction and a vent to allow for chamber
pressure equilibration. Tubing was attached to the vent prior to capping and hung inside the
chamber to reduce possible crosswind induced loss of gas from the chamber vent. Headspace
gas from within the chambers was extracted immediately following lid placement with a 30 mL
nylon syringe and a 23-gauge needle. Three subsequent extractions were made at 20-min
intervals over a 60-min period. Glass 5.9 mL Exetainer vials (Labco Limited, Buckinghamshire,
UK) were flushed with 20 mL of extracted sample and then overcharged with 10 mL of sample
to facilitate sample extraction for analysis (Parkin & Venterea, 2010). Field standards (1 ppm
N2O) and ambient air were also loaded into vials at this time to assess ambient GHG
concentrations and potential storage-vial degradation in the period between sampling and
analysis. Samples were analyzed by gas chromatography using an electron capture detector
(micro-ECD, Agilent 7890A GC System, Santa Clara, CA, USA).
“Carbon dioxide (CO2) accumulation was used to evacuate compromised vial through a
visual inspection process which could result in deletion of a single observation in a series, or the
removal of the entire series. Samples passing visual inspection were analyzed with the HMR
package (v0.3.1, Pedersen 2012) in the R statistical environment (“warm puppy” version, R Core

26

Team, 2013). Briefly, the method fits trace gas concentration time series with either a nonlinear
model (Hutchinson and Mosier 1981) or linear regression, or identifies the series as a null flux.
When the 95% confidence interval of a nonlinear flux estimate did not include the
corresponding linear flux estimate, the nonlinear estimate was used for that series; in all other
cases, linear flux estimates were used. Daily fluxes were aggregated to an annual scale by
linearly interpolating between consecutive sampling dates and integrating (Smith & Dobbie,
2001).”
A decision making tree was constructed by the Area 4.3 research group to determine
whether a given flux was linear, non-linear, or absent. If a non-linear flux were assumed to be
linear, the nitrous oxide emission would be greatly overestimated for that sample. Non-linear
fluxes were quantified using the Hutchinson and Mosier (1981) approximation. Absent fluxes
were given a value of ‘zero’. Once the daily fluxes were determined, the annual nitrous oxide
flux was calculated for each system by linear interpolation. The annual nitrous oxides fluxes for
2008 through 2012 were averaged into an average annual N2O flux value for each system.

Gas Flux and Models
Nitrous oxide emissions were considered non-material outputs in the cropping system
models. Values were plugged into the cultivation process of each model and are located in
Table 16 of the Appendix for the ARL cropping system models and in Table 10 of the Appendix
for the KBS cropping system models.
Although carbon dioxide flux data has been collected by the Area 4.3 research group for
the cropping systems at both locations, the short term duration of the study was considered
insufficient to make meaningful conclusions on soil carbon changes attributable to the specific
27

cropping systems. Therefore, carbon dioxide flux values were based on 20-year soil carbon
research conducted at the Long Term Ecological Research (LTER) experiment located at Kellogg
Biological Station (Gelfand, in press). If soil carbon is increasing for a treatment, the treatment
is sequestering carbon in the form of carbon dioxide from the atmosphere. This will decrease
the global warming potential for the treatment. If soil carbon is decreasing for a treatment, the
treatment is emitting carbon in the form of carbon dioxide to the atmosphere. This will increase
the global warming potential for the treatment. Carbon dioxide values are considered nonmaterial inputs (sequestration) or non-material outputs (emission) in the cropping system
models. Carbon dioxide flux values are plugged into the cultivation process of each model and
are located in Table 13 of the Appendix for the ARL cropping system models and in Table 13 of
the Appendix for the KBS cropping system models.

GaBi 6.0 Professional + Extension Database
Aggregated process data contained in the Professional + Extension 2012 database in
GaBi 6.0 was used for the material and non-material input extraction and production processes
in the cropping system models. Processes derived from US data were used whenever possible
in the cropping system models. When US derived processes were absent from the Professional
+ Extension 2012 database, processes derived from Canadian data were used instead. If both US
and Canadian derived processes were absent from the database, any available European
processes were used. Cropping system models differ by location due to differences in
agronomic inputs and management practices. The final ARL cropping systems models are
depicted in Figures 4, 5, 6, 7, 8, 9, 10, 11, 12, and 13. The final KBS cropping system models are
depicted in Figures 14, 15, 16, 17, 18, 19, 20, 21, 22,and 23. Solid blue arrows indicate material
28

inputs. Solid black arrows indicate material energy inputs. Dashed blue arrows indicate nonmaterial inputs. Arrows are scaled to show their relative contribution to the system based on
mass in kilograms. Material and non-material outputs are not depicted in the cropping system
models.

Figure 4. Cropping system model of treatment G1 Continuous at ARL.

29

Figure 5. Cropping system model of corn in rotation (corn
(corn-rot) at ARL.

(soybean-rot) at ARL.
Figure 6. Cropping system model of soybeans in rotation (soybean

30

Figure 7. Cropping system model of canola in rotation (canola-rot) at ARL.

31

Figure 8.. Cropping system model of normalized corn
corn-soybean-canola
canola rotation (rotational
system) at ARL.

Figure 9. Cropping system model of G5 Switchgrass at ARL.
32

Figure 10. Cropping system model of G6 Miscanthus at ARL.

Figure 11. Cropping system model of G7 Native Grass mix at ARL.

33

Figure 12. Cropping system model of G9 Old Field at ARL.

Figure 13. Cropping system model of G10 Native Prairie mix at ARL.

34

Figure 14. Cropping system model of treatment G1 Continuous at KBS.

Figure 15. Cropping system
ystem model of corn in rotation (corn
(corn-rot) at KBS.

35

Figure 16. Cropping system model of soybeans in rotation (soybean
(soybean-rot)
rot) at KBS.

Figure 17. Cropping system model of canola in rotation (canola
(canola-rot) at KBS.

36

Figure 18. Cropping system model of normalized corn
corn-soybean-canola
canola rotation (rotational
system) at KBS.

Figure 19. Cropping system model of G5 Switchgrass at KBS.

37

Figure 20. Cropping system model of G6 Miscanthus at KBS.

Figure 21. Cropping system model of G7 Native Grass mix at KBS.

38

Figure 22. Cropping system model of G9 Old Field at KBS.

Figure 23. Cropping system model of G10 Native Prairie mix at KBS.

Life-Cycle Impact Assessment (LCIA)
The purpose of the life cycle impact assessment phase is to characterize LCI results into
impact categories consistent with the LCA goal. Impact categories can be classified as either
midpoint or endpoint. Midpoint categories reflect the magnitude and significance of an
environmental emission (calculated in the life cycle inventory) for a system (Bare, 2002).

39

Midpoint impact categories require simpler models and reduce the uncertainty of the LCIA
results due to limiting forecasting and assumptions as compared to endpoint impact categories.
Endpoint categories reflect the larger impacts of a life cycle by aggregating multiple impact
categories into a single value or score. When aggregating impact categories, specific categories
can be weighted based on their perceived relevance to the endpoint. Endpoint impacts can be
more comprehensive and leave less room for interpretation than midpoint impacts. The
advantages and disadvantages of both types of impact categories should be considered when
selecting a life cycle impact assessment methodology.
The goal of this life cycle assessment research was to determine the global warming
potential (GWP) of each cropping system, therefore a life cycle impact assessment (LCIA)
methodology that included GWP was required for this study. When choosing an LCIA
methodology it is also important to consider the geographic parameters included in the
methodology framework. Many of the original LCIA methodologies were developed with
European parameters and considerations, for example CML 2001 and ReCiPe were both
developed in the Netherlands. The geographic scope of this LCA was the north central United
States, so methodologies developed for use in Europe likely would not accurately reflect
environmental impacts or systems in the US. The LCIA methodology used in this study was
TRACI 2.0 because it includes GWP and was developed for use in the US.

TRACI Impact Assessment Methodology
The Tool for the Reduction and Assessment of Chemical and other environmental
Impacts (TRACI) impact assessment methodology was developed by the US Environmental
Protection Agency (EPA) in 2002 with US parameters and characterization factors (Bare, 2002).
40

An updated and expanded version, TRACI 2.0, was released in 2011. TRACI 2.0 methodology is
built into the LCA software, GaBi 6.0, used in this study.
TRACI 2.0 includes midpoint impact categories for ozone depletion, global warming,
acidification, eutrophication, smog formation, human health (particulate, cancer, non-cancer),
ecotoxicity, and fossil fuel use (Bare, 2002). Figure 24 shows the impact categories included in
TRACI 2.0 and their possible endpoints.

Figure 24. An overview of TRACI 2.0 midpoint impact categories and possible endpoints.

41

Global Warming Potential Impact Characterization
Greenhouse gases (GHG) that contribute to global warming potential (GWP), as
determined by Intergovernmental Panel on Climate Change (IPCC), are included in the TRACI
2.0 midpoint category for GWP. Each gas has its own characterization factor based on a
chemical’s radiative efficiency and lifetime (IPCC, 2007) as compared to carbon dioxide (the
established measurement system). Characterization factors for GWP are determined for a given
time horizon, either 20-year, 100-year, or 500-year (IPCC, 2007). This study will use the 100year time horizon. The mass of an emission is multiplied by the characterization factor resulting
in the global warming potential for that emission. The global warming potential for every
emission of the life cycle is summed to get the total global warming potential (Bare, 2002).
Global Warming PotenƟal of the life cycle = ෍ e௜ ൈ GWP௜
௜

e௜ =emission in kg
GWP௜ ൌ global warming potenƟal of substance ݅=characterizaƟon factor of substance ݅
The GHG emissions from each system were calculated in the life cycle inventory phase.
The pathway between the life cycle inventory phase and the life cycle impact assessment phase
for GWP is shown in Figure 25.

42

Figure 25. Global warming potential pathway starting at the life cycle inventory phase.

RESULTS
Life Cycle Interpretation
Global Warming Potential
Nitrous Oxide Emissions and Global Warming Potential
The nitrous oxide emissions of a cropping system had a large impact on the global
warming potential (GWP) of that system. The average annual nitrous oxide emissions are
shown in Figure 26. Error bars represent one standard deviation of the treatment mean for the
annual nitrous oxide emissions of each system.

43

Figure 26. Average annual nitrous oxide emissions of each cropping system for the study years
2008-2012. The * indicates the rotational system values were composed of corn-rot,
corn
soybeanrot, and canola-rot data and not directly measured. Error bars represent one standard deviation
of the treatment mean between the annual nitrous oxide measurements for each system.

An independent-samples
samples Student’s t-test was conducted to compare the nitrous oxide
emissions at ARL and KBS for each system. Welch’s t-test was used for the soybean-rot
soybean
and G10
t-tests to account for unequal variances between locations identified from the F-test (data not
shown). All other cropping systems were assumed to have equal variances between locations
based on the results of the F-test
test (data not shown). The research hypothesis was that the
nitrous oxide emissions for a system would differ by location; the null hypothesis was that the
nitrous oxide emissions for a system would be the same at both locations.

44

Table 4. Results of the Student’s t-test for testing the difference in nitrous oxide emissions
between locations for each cropping system (α=0.10).
System

ARL mean

KBS mean

p-value

G1 Continuous Corn

7.58

8.62

0.8225

Corn-rot

8.35

6.21

0.5996

Soybean-rot

4.15

1.46

0.1078

Canola-rot

4.20

2.64

0.4370

Rotational System

5.84

3.74

0.2326

G5 Switchgrass

6.32

2.41

0.1731

G6 Miscanthus

3.44

3.03

0.7159

G7 Native Grass Mix

4.06

3.34

0.8127

G9 Old Field

2.93

2.00

0.2551

G10 Prairie Grass Mix

4.11

1.44

0.4524

The results of the two-tailed t-test failed to reject the null hypothesis for all systems
(Table 4.) at the 90% confidence level (α=0.10). Therefore, the nitrous oxide emissions for a
system did not differ significantly by location.
Analysis of variance, ANOVA, was also conducted on the nitrous oxide emissions at each
location. The ANOVA was used to test whether the means of nitrous oxide emissions were
equal for all cropping systems at one location. The research hypothesis was that cropping
systems did not emit equal amounts of N2O; the null hypothesis was that the cropping systems
emitted equal amounts of N2O. At ARL, the results of the ANOVA failed to reject the null
hypothesis (p=0.6170, α=0.10). At KBS, the results of the ANOVA rejected the null hypothesis
(p=0.0779, α=0.10) meaning that nitrous oxide emissions were not equal across all treatments
at this location. Because the results of the ANOVA were not significant at ARL, further post-hoc
45

testing (multiple comparisons) was not performed. Multiple comparison testing was performed
for KBS to identify the treatments with significant differences in nitrous oxide emission. At
α=0.10, the N2O emissions from the G1 continuous corn system and the G6 miscanthus system
were greater than the N2O emissions from soybean phase of the rotational system, p=0.0969
and p=0.0573 respectively.
The yearly variation in N2O emissions for the perennial systems is shown in Figure 27.
Perennial systems at ARL had relatively high initial emissions in 2008 compared to the other
study years. The perennial systems at KBS did not exhibit the same initial pulse of emissions as
ARL; they remained fairly constant across study years.

46

Figure 27. Nitrous oxide emissions by year for the perennial cropping systems at each location.
N2O emission means within each treatment labeled with the same letter are not significantly
different from each other (t-test,
test, α=0.10).

The yearly variation in N2O emissions for the annual systems is shown in Figure 28. Both
locations seemed to experience large pulses in nitrous oxide emissions in 2010. Annual systems
at ARL in 2012 also saw a moderate pulse in emissions relative to the previous year.

47

Figure 28. Nitrous oxide emissions by year for the annual cropping systems at each location.
N2O emission means within each treatment labeled with the same letter are not significantly
different from each other (t-test,
test, α=0.10). 2012 data was not available for KBS.

Material Inputs and Global Warming Potential
The material inputs that contributed most to the global warming potential of each
cropping system were nitrogen fertilizers. Figure 29 and figure 30 show that urea ammonium
nitrate (28% UAN) and ammonium nitrate contribute the most to the GWP of each system out
of all of the material inputs.

48

Figure 29. Material input contribution to the global warming potential of each system at
Arlington Agricultural Research Station (ARL). Golden yellow bars represent urea ammonium
nitrate and royal blue bars represent ammonium nitrate.

Figure 30. Material input contribution to the global warming potential of each system at
Kellogg Biological Station (KBS). Golden yellow bars represent urea ammonium nitrate, grey
bars represent urea, and orange bars represent monoammonium phosphate.

49

ARL used urea ammonium nitrate (UAN) for corn cropping systems and ammonium
nitrate for the canola and perennial cropping systems that were fertilized. The differences in
nitrogen fertilizer type had an impact on GWP. Ammonium nitrate contributed 9.15 kg CO2-eq
to the GWP per pound of nitrogen (N) applied. Urea ammonium nitrate, a mixture of urea and
ammonium nitrate, contributed 6.5 kg CO2-eq to the GWP per pound of N applied. The use of
ammonium nitrate instead of UAN lead to a 28.9% increase in fertilizer contributions to GWP
for the perennial and canola systems. Fertilizer production effects were most likely the cause
for this difference. Fertilizer production data for ammonium nitrate and UAN is contained in the
Professional + Extension 2012 database and was derived from US production sources.
KBS used UAN as the main nitrogen source in all fertilized cropping systems and
contributed the majority of GWP from material inputs for each system. Corn systems also have
a sizable GWP contribution from monoammonium phosphate (MAP). MAP is commonly used as
a starter fertilizer in corn because it supplies both nitrogen and phosphorous, two of the
essential nutrients for plants. The canola phase of the rotation system received urea one year
instead of UAN. The use of urea instead of UAN lead to a 41.5% decrease in fertilizer
contributions to GWP for the year urea was used compared to the years UAN was used in the
canola system.

Global Warming Potential on a Land Area Basis
Figure 31 shows the global warming potentials (GWP) for each cropping system in the
GLBRC Biofuels Cropping System on a land area basis.

50

Figure 31. Average global warming potential of each cropping system on a land area basis per
year for the study years 2008-2012
2012 (GWP kg CO2-equiv ha-1 yr-1).

Recall, the functional unit for the LCA was one hectare of arable land per year. Except
for treatment G1 - continuous corn, ARL cropping systems had a higher GWP than KBS cropping
systems. A negative global warming potential indicates a cropping system that is sequestering
more GHG from the atmosphere than it is emitting.

Yield
Although GWP is greatest for the corn cropping systems, these systems have the highest
potential biofuel yields under current technologies. The grain yield of annual systems and the
dry biomass yield of perennial systems were recorded for each location annually (Figure 32).

51

Figure 32. Annual grain yield of annual systems and dry biomass yield of perennial systems.
Means within each treatment labeled with the same letter are not significantly different from
each other (t-test, α=0.05). Yield unit is Mg ha-1.

Starch-based ethanol conversion was assumed for corn grain systems, biodiesel
conversion for soybean and canola systems, and cellulosic ethanol conversion for perennial
systems. Average energy (MJ) produced by each system on one hectare of cropping
cropp
system is
shown in Figure 33.

52

Figure 33. Average estimated energy (MJ) produced on one hectare of each cropping system
per year for study years 2008-2012
2012. Energy estimates based on grain ethanol for corn systems,
biodiesel for soybean and canola systems, and cellulosic ethanol for perennial systems.
Cropping system means labeled with the same letter are not significantly different between
treatments per location. (t-test, α=0.10).

Conversion estimates from dry biomass to ethanol do not yet exist for the specific plant
communities contained in G6, G7, G9, and G10, so switchgrass conversion factors and estimates
were assumed for these systems.
The global warming potential of the estimated amount of gasoline displaced by each
cropping system was calculated (Figure 34.). Gasoline displacement calculations assumed the
energy in one unit of ethanol (E100) was equal to 0.7 units of gasoline (ethanol producing
systems) and one unit of biodiesel (B100) was equal to 0.921 units of petroleum-based
petroleum
diesel
(biodiesel producing systems). Gasoline and diesel were assigned a GWP of 91 g CO2-eq per MJ
of energy based on the standard EPA estimates in 2005.

53

Figure 34. Net global warming potential mitigated (blue bar) by the biofuel crop when
compared to equal amounts of fossil fuel energy produced by each system.

An analysis was also performed on the global warming potential per MJ of energy
produced in each system. Results are shown in Figure 35. Negative values suggest that the
system is sequestering more GHG from the atmosphere than it is emitting during the cropping
phase of its life cycle. Theoretically, favorable bioenergy cropping systems should appear in the
upper left portion of the graph (high energy output and negative net GWP). The cropping

54

systems containing corn are generally the closest to this location mainly due to high yields, and
therefore potential high ethanol production.

Figure 35. Net global warming potential (kg CO2-eq ha-1 yr-1) verses the energy produced, in MJ,
by each cropping system. Cropping systems are labeled above their respective data point.

DISCUSSION
Nitrous Oxide Emissions
Nitrous Oxide emissions had the biggest effect on the global warming potential for a
system due to the radiative forcing of N2O in the atmosphere. Nitrous oxide is produced
naturally in agricultural systems through the processes of nitrification and denitrification
(Lemke et al., 2008). When weather conditions are warm and moist, soil microbes convert
ammonium (NH4+) to nitrate (NO3-). If nitrate is not taken up by plants it can either be leached
from the soil, or denitrified if soil conditions and microbial activity change (Lemke et al., 2008).
Denitrification is the microbial process of converting nitrate into N2O or N2 gases. The gases are
55

released as emissions from the soil into the atmosphere. N2 gas does not have a radiative effect
in the atmosphere, and therefore is not considered a greenhouse gas. Denitrification occurs in
saturated soils (Lemke et al., 2008). Soils can become saturated after periods of heavy rainfall
and from snow melt in the spring. Nitrous oxide emissions are generally higher during the
spring due to frequent rain events and snowmelt.
The large variations in annual nitrous oxide emissions can explain why no significant
differences were identified between locations for each cropping system (Student’s t-test) or
between cropping systems at each location (ANOVA). The variations in annual N2O emissions
could be due to a number of reasons including local climate conditions and soil type and SOM.
ARL experienced slightly cooler temperatures and slightly higher annual precipitation than KBS.
The increased precipitation could have increased the soil moisture at ARL and resulted in a
greater amount of denitification compared to KBS (Lemke et al., 2008). ARL soils also had a
higher SOM content than KBS soils, 3.8% and 2.4% respectively. The decomposition of organic
matter could increase the N2O emissions from high SOM soils (Toma et al., 2011). Although the
climate and SOM contents suggest ARL should have higher average annual nitrous oxide
emissions, the statistical analysis suggests this is not the case due to high variability in nitrous
oxide emissions across both locations.

Fertilizer Sources
The results of the LCA revealed that differences in nitrogen fertilizer source had an impact
on the global warming potential. Ammonium nitrate (AN) (34-0-0) contributed about 9.15 kg CO2eq per lbs of N applied; UAN (28-0-0) contributed about 6.5 kg CO2-eq per lbs of N applied; and
urea (46-0-0) contributed about 2.69 kg CO2-eq per lbs of N applied. 28% UAN is a mixture of equal
56

parts urea and AN with enough water to keep the solution in suspension and make up the
remainder of the volume. Differences in nitrogen fertilizer source GWP is most likely due to the
production phase of the fertilizers. The production phase of a fertilizer includes the extraction of
raw materials, the production of energy required throughout the fertilizer production process, and
the transportation of fertilizers to their destination (Skowroñska et al., 2014). Fertilizer production
mainly contributes to GWP by CO2 emissions from fossil fuels used in ammonia production and by
N2O emissions during the production of nitric acid. Overall GWP contributions will depend on the
type of fertilizer. A recent literature review on fertilizer LCAs by Skowroñska et al., showed similar
differences between AN and urea contributions to GWP as this LCA. According to Skowroñska et
al., the production phase of AN contributes 6.2 kg CO2-eq per kg of AN produced; urea contributes
1.59 kg CO2-eq per kg of urea produced. Although the scales and system boundaries of the
Skowroñska et al. are slightly different from those of this study, the same general trends in
fertilizer GWP can be seen. Due to lower GWP from the production of urea fertilizers, future
fertilizer recommendations could possibly highlight urea as a more ‘environmentally friendly’
fertilizer choice. It is important to point out that urea fertilizers have the potential to be highly
volatile if not incorporated into the soil (by tillage or rainfall) soon after application. Urease
inhibitors are also a viable option to decrease the volatilization of urea fertilizers.

Corn Stover
Corn stover was not considered in this LCA because stover removal for ethanol
production was not the main plot treatment in the continuous corn system. However, because
of future large-scale production potential of cellulosic ethanol from corn stover it should be
addressed. Corn stover was harvested in the micro-plot section of each corn treatment every
57

year and yields were recorded. Corn stover was removed at a rate at which the soil organic
matter content of the soil would not be greatly affected. Removing too much stover can cause a
net loss of carbon from the field resulting in higher erosion, reduced soil tilth, and a net
emission of carbon to the atmosphere (likely in the form of CO2, a GHG). ARL soils have a higher
amount of SOM compared to KBS soils, therefore larger amounts of stover were able to be
removed at ARL than KBS. The stover harvest at ARL averaged 5.84 Mg ha-1 yr-1. Based on a
conversion rate of 80 gallons of ethanol per ton (Sheaffer et al., 2010), the stover has the
potential to produce 41384 MJ ha-1 in addition to the 99522.12 MJ ha-1 from corn grain ethanol
in the same system. At KBS, the average annual stover harvest was 2.07 Mg ha-1 yr-1, capable of
producing an additional 14687 MJ ha-1 for this system.

Perennial Crops
Miscanthus
Based on yield, miscanthus appears to be a promising perennial crop due to its high
biomass yields and therefore high cellulosic ethanol yield potential. The miscanthus cropping
system performed well in the life cycle analysis. The global warming potential for miscanthus was
slightly positive at ARL (68.7 kg CO2-eq ha-1 yr-1), but negative at KBS (-354 kg CO2-eq ha-1 yr-1).
Due to an unadvised first-year biomass harvest, miscanthus winter killed after the 2008 growing
season at ARL. The re-establishment of the miscanthus stand at this location may have
contributed to a slightly positive GWP. Biomass was not harvested in 2009 or 2010 at ARL to
avoid winter kill in the re-established stand. The lack of yield data for these years reduced the dry
biomass yield average for ARL miscanthus as seen in Figure 33. If miscanthus had not winter killed
at ARL the average yield would be expected to exceed the KBS average yield for 2008-2012.
58

Old Field Treatment
The biomass yields for treatment G9 - Old Field were fairly consistent across study years
for each location. However, the species composition of the harvested biomass changed
drastically. At ARL, the plant communities in 2008 and 2009 were dominated by summer annual
species: Echinochloa crus-galli (L.) beauv (barnyard grass), Setaria pumila (Poir.) Roem & Schult
(yellow foxtail), Abutilon theophrasti (velvetleaf), Chenopodium album (L.) (lambsquarters), and
Polygonum pensylvanicum (L.) (Pennsylvania smartweed). In 2010, the plant communities
began to shift to a mixture of summer and winter annuals dominated by: Conyza Canadensis
(L.) cronq. (marestail) and Lactuca serriola (L.) (prickly lettuce). In 2011, the plant community
was dominated completely dominated by winter annuals and biennials: Carduus acanthoides
(L.) (plumeless thistle), prickly lettuce, and Cirsium vulgare (Savi.) Tenore (spear thistle). Carson
et al. (1988), also identified this successional plant community shift from summer annuals to
winter annuals and perennial by the third year of the treatment. Originally, the dominant
species were weeds well suited for the disturbed environments of production agricultural
systems. Over the study years, increasing competition for resources including light, water, and
nutrients resulted in a shift to a competition based dominant species community composition
(Lohbeck et al, 2014).
The KBS G9 - Old Field treatment did not start to transition into a winter annual and
perennial plant community as described by Carson et al. (1988) until 2012. Species composition
was unavailable for 2008. In 2009, 2010, and 2011, the plant community was dominated by
summer annual species: lambsquarters, Digitaria sanguinalis (L.) scop (large crabgrass),
barnyard grass, marestail, and Setaria viridis (L.) Beauv. (green foxtail). In 2012, the plant

59

community was dominated by the summer annuals marestail and Setaria faberi Herrm.
(Japanese bristlegrass), and the winter annual prickly lettuce. It would be interesting see if the
2013 plant community at KBS was completely dominated by winter annuals and perennials or if
2013 was another transition year. The slower transition at KBS could have been due to the
marginally productive soils compared to the highly productive soils as ARL. Less productive soils
would likely have increased competition for resources within a plant community. With more
competition for limited resources, the initial plant species would dominate for a longer period
due to the fact that they were already established and reproducing and effectively smothering
out other emerging plant species (the winter annuals and perennials).

Land Use Changes
This experiment did not evaluate the potential land use changes for converting prime
agricultural land into land for biofuel feedstock production. However, it is worth noting that the
perennial cropping systems (G7, G9, G10) with the lowest GWP on a land area basis also
produced the least amount of energy per hectare. At ARL, it would take five-times the amount
of land under the unfertilized prairie grass mix cropping system to produce the same amount of
energy as the continuous corn cropping system per year. At KBS, that factor increases to seventimes the amount of land under the unfertilized prairie grass mix cropping system to produce
the same amount of energy as the continuous corn cropping system per year.
As the world’s population grows, food production will also have to increase. Converting
hectares of prime farmland to biofuel feedstock production isn’t reasonable or sustainable for
human life. If bioenergy policy continues to push for annual biofuel production requirements,
alternative land must be considered. Currently, research is under way on the use of ‘marginal
60

lands’ for biofuel feedstock production. Marginal land is land that is not fit for crop production
due any number of reasons including climate, soil type, water availability, soil fertility, and/or
government land management programs. Due to the marginality of this land, successfully
establishing perennial crops could be challenging and costly. In addition, fertilizer and soil
amendment requirements could be high in order to grow a harvestable quantity of biofuel
feedstock. And throughout all of this, economics would be a driving factor in whether or not
biofuel crops would be accepted by farmers.

Model Assumption Effects and Considerations
Due to many the assumptions made during this life cycle assessment study, a cautious
approach should be taken when interpreting the results. One of the biggest limitations of this
study that lead to numerous assumptions was the GaBi LCA software and Professional +
Extension 2012 database. GaBi 6.0 is a powerful LCA program. After several classes and
tutorials on the software, the practitioner of this study was still only able to perform basic to
intermediate level operations and analysis with the software. In addition, the amount of
agricultural inputs and processes included in the Professional + Extension 2012 database
included in the software was severely limited. Given the nature of the GLBRC Biofuels Cropping
Experiment, many of the cropping systems are novel (the perennial systems) and therefore lack
the substantial amounts of peer-reviewed literature that is required for inclusion in standard
databases.

61

Miscanthus Rhizomes
There was not a process in the Professional + Extension 2012 database for Miscanthus
rhizomes. In fact, there were not any agricultural processes that even remotely resembled or
that could be used to represent a miscanthus rhizome. Due to the lack of adequate alternatives,
rhizomes as a material input for the miscanthus cropping system were not included in this
study. A relatively huge number of rhizomes were used in the field experiment (17,000+
rhizomes ha-1), therefore it can be assumed that the inclusion of rhizomes would impact the
results of the miscanthus cropping system in this study. Because all of the other crop seed
processes in this study contributed negative values to global warming potential (decreased
GWP), the inclusion of miscanthus rhizomes could potentially have the same effect on the
miscanthus cropping system GWP.
The current GWP values for this miscanthus cropping systems at ARL and KBS are 68.7
and -354 kg CO2-equiv ha-1 yr-1 respectively. If these GWP values were adjusted based on the
seed input values for the other cropping systems in the experiment the following new values
could be considered for miscanthus:
At ARL, the seed input for the cropping systems accounted for about 1.36% of the total
cropping system GWP with the exception of the soybean system at 38.59%. Based on the other
cropping system seed input values, if miscanthus rhizomes reduced the currently modeled
system’s total GWP by 1.36% to 38.59%, the resulting GWP range would be 36.42 to 59.31 kg
CO2-equiv ha-1 yr-1.
At KBS, the seed input for the cropping systems accounted for about 1.65% of the total
cropping system GWP with the exception of the soybean system at 29.85%. Based on these

62

seed input values, if miscanthus rhizomes reduced the currently modeled system’s total GWP
by 1.65 to 29.85%, the resulting GWP range would be -412.58 to 535.35 kg CO2-equiv ha-1 yr-1.

Crop Residues as a Co-Product
GaBi 6.0 is a LCA program, not an agricultural systems modeling program. It is designed
to model static systems with consistent inputs and outputs. Agricultural systems are the
complete opposite of a static system. Agricultural systems are dynamic, always changing
depending on number of biotic and abiotic factors at any given time. The program’s inability to
account for soil dynamics, climate influences, and even field residues as co-products limited this
LCA study.

CONCLUSION
A life cycle assessment was performed on each cropping system in the GLBRC Biofuels
Cropping System Experiment to identify the global warming potential of one hectare of arable
land of each system. Corn cropping systems consistently had the highest GWP at both locations.
The unfertilized prairie grass system, G10, had the lowest GWP at both locations. The GWP of a
cropping system was heavily influenced by the nitrous oxide emissions for that system. A
statistical analysis revealed no significant differences between locations for each cropping
system (Student’s t-test) or between cropping systems at each location (ANOVA) likely due to
the large variations in annual nitrous oxide emissions. G6 - miscanthus, a perennial biomass
crop, showed the greatest potential in terms of low GWP and high estimated cellulosic ethanol
yield, and therefore the best net GWP from the large displacement of fossil fuels. Due to the

63

exclusion of miscanthus rhizomes as a material input, this particular phase of the life cycle
assessment should be investigated in greater detail before a large number of resources are
contributed to the development of miscanthus as a biofuel feedstock based on the conclusions
of this study.
Limitations imposed by the LCA software and database lead to many model
assumptions in this study. These limitations and assumptions require a conservative
interpretation of study results.
Potential land use changes, not considered in this study, could have a major impact on
future bioenergy policy and biofuel requirements. Potential land use changes and the feasibility
of growing dedicated bioenergy crops on marginal lands should be a major area of focus for
future biofuel research.

64

APPENDIX

65

APPENDIX

Climatological Data
Annual Temperature and Precipitation

Figure 36. Annual temperature and precipitation for Arlington Agricultural Research Station
(ARL) and Kellogg Biological Station (KBS). Annual and 30-year average weather data from
National Climatic Data Center (NCDC). 30-year average weather data is from the latest NCDC
Climate Normals report for 1981-2010.

66

Growing Season Temperature and Precipitation
Kellogg Biological Station

Figure 37. Yearly air temperature and precipitation data from the Hickory Corners weather
station recorded by Michigan State University Enviro
Enviro-weather service. Kellogg Biological
Station is located in Hickory Corners, Michigan in close proximity to the weather station.
Weather data for 07/15/2009-08/31/2009
08/31/2009 was obtained from the Kalamazoo weather station
due to no reported data from the Hickory Corners weather station during this timeframe.
Kalamazoo is located approximately 18 miles so
south-east of Hickory Corners.

67

Figure 37 (cont’d)

68

Arlington Agricultural Research Station

Figure 38. Yearly air temperature and precipitation data from the Arlington University Farm
weather station recorded by the National Weather Service.

69

Figure 38 (cont’d)

70

Variety and Species Composition
Perennial Crop Mixes
Table 5. Species planted in the perennial cropping systems. Species are the same at both study
locations. C4 and C3 indicates the photosynthetic pathways of particular grass species.
System

System Name

Group

G5

Switchgrass

Graminoids

G6

Miscanthus

G7

Native Grass Mix

G10

Prairie Grass Mix

Species (latin)

Common Name / Variety

C4

Panicum virgatum

switchgrass / Cave-in-rock

Graminoids

C4

Miscanthus x giganteus

giant miscanthus

Graminoids

C3

Elymus canadensis

Canada wildrye

C4

Panicum virgatum

switchgrass

Andropogon gerardii

big bluestem

Schizachyrium scoparium

little bluestem

Sorghastrum nutans

Indiangrass

Koeleria cristata

prairie Junegrass

Elymus canadensis

Canada wildrye

Panicum virgatum

switchgrass

Andropogon gerardii

big bluestem

Schizachyrium scoparium

little bluestem

Sorghastrum nutans

Indiangrass

Desmodium canadense

showy ticktrefoil

Lespedeza capitata

roundhead lespedeza

Baptisia leucantha

white false indigo

Rudbeckia hirta

blackeyed Susan

Anemone canadensis

Canadian anemone

Asclepias tuberose

butterfly milkweed

Silphium perfoliatum

cup plant

Monarda fistulosa

wild bergamot

Ratibida pinnata

pinnate prairie coneflower

Solidago rigida

rigid goldenrod

Solidago speciosa

showy goldenrod

Aster novae-angliae

New England aster

Graminoids

C3

C4

Legumes

Early forbs

Mid forbs

Late forbs

71

Annual Crops
Kellogg Biological Station
Table 6. Annual crop variety by year at KBS.
System

Year

Variety

G1

2008

Dekalb DKC 52-59 Corn Hybrid

Notes

2009
2010
2011
2012
corn-rot

2008

Dekalb DKC 52-59 Corn Hybrid

2009
2010
2011

soybean-rot

2012

Dekalb DKC 48-12 RIB Corn Hybrid

2008

Pioneer 91M80 RR Soybeans

2009

Pioneer 92Y30 Soybeans

2010
2011
2012
canola-rot

2008

winter canola; Dekalb DKL 52-10 RR Spring Canola

2009

Dekalb DKL 52-10 RR Spring Canola

2010
2011

Dekalb Genuity RR Canola DKL 72-55

72

no variety information for winter
canola; canola was replanted in
spring

Arlington Agricultural Research Station
Table 7. Annual crop variety by year at ARL.
System

Year

Variety

G1

2008

Dekalb 52-59 RR

2009

Pioneer 35F40 RR/LL/Bt

2010

Delkalb 52-59 RR/Bt

Notes

2011

corn-rot

2012

Pioneer 35F40

2008

Dekalb 52-59 RR

2009

Pioneer 35Y40 RR/LL/Bt

2010

Dekalb 52-59 RR/Bt

2011

soybean-rot

canola-rot

2012

Great Lakes 4041G3VT3

2008

Pioneer 92Y40 RR

2009

Dairyland 2200 RR

2010

Pioneer 92Y51

2011

Dairyland 2375/R2Y

2012

Dairyland 2011

2008

Dekalb 5210 RR

2009

InVigor 8440 LL

spring canola

2010
2011

73

Herbicide
Arlington Agricultural Research Station
Table 8. Herbicides used at ARL. Chemical and trade names are listed.
Chemical Name

Trade Name(s)

Crops

glyphosate

Honcho Plus; Mirage Plus; Roundup
PowerMAX

corn; canola; switchgrass;
miscanthus; native grass mix;
prairie grass mix

s-metolachlor

Dual II Magnum;

corn; soybean; miscanthus

2,4-D ester

corn; soybean; canola;
switchgrass; miscanthus; native
grass mix

mesotrione

Callisto

corn

simazine

Princep

corn

tembotrione

Laudis

corn

sulfentrazone; cloransulam-methyl

Authority First

soybean

glufosinate
dicamba

soybean; canola
Clarity

switchgrass

quinclorac
pendimethalin

switchgrass
Prowl

miscanthus

74

Kellogg Biological Station
Table 9. Herbicides used at KBS. Chemical and trade names are listed.
Chemical Name

Trade Name(s)

Crops

glyphosate

Roundup OriginalMAX; Roundup
PowerMAX;

corn; soybean; canola

s-metolachlor; atrazine; mesotrione Lexar

corn

2,4-D ester

corn; soybean

glufosinate

Ignite

corn

quinclorac

Drive

switchgrass; miscanthus

2,4-D amine

switchgrass

75

Cropping System Model Data
Kellogg Biological Station
Material Inputs
Table 10. Material inputs at KBS per reference flow of 1 ha year-1.

76

Non-Material Inputs
Table 11. Non-material
material inputs at KBS per reference flow of 1 ha year-1.

77

Material and Energetic Output
Table 12. KBS: Yield refers to grain yield at standard moisture content for corn, soybeans,
canola; dry biomass yield of perennial crops. Per reference flow 1 ha year-1.
-1

-1

System

Yield (Mg ha )

Potential Energy (MJ ha )

Source

G1

9.16

74780.67

KBS field data corn ethanol energy [1]

corn-rot

8.15

66530.07

corn ethanol energy [1]

soybean-rot 2.67

18500.86

soybean biodiesel energy [2]

canola-rot

1.18

18919.14

canola biodiesel energy [3]

G5

2.98

19056.40

equation [4]

G6

8.65

55199.40

equation [4]

G7

1.98

12667.46

equation [4]

G9

2.17

13876.94

equation [4]

G10

1.67

10682.08

equation [4]

78

Notes

Soil Carbon Accumulation
Table 13. KBS: Carbon dioxide accumulation in the soil for each cropping system. Negative
values indicate a net mitigation. Source: Gelfand et al., 2014.

Nitrous Oxide Emissions
Table 14. KBS: Nitrous oxide emissions from each cropping system. Source: 4.3, in press.
-1

-1

-1

Cropping System

kg N2O-N
N ha yr

kg N2O ha yr

G1 Continuous Corn

5.4849

8.6191

corn-rot

3.9521

6.2104

soybean-rot

0.9334

1.4668

canola-rot

1.6856

2.6487

rotational system

2.3792

3.7388

G5 Switchgrass

1.5362

2.4140

G6 Miscanthus

1.9284

3.0304

G7 Native Grass mix

2.1031

3.3047

G9 Old Field

1.2748

2.0033

G10 Prairie Grass mix

0.8962

1.4082

79

-1

Notes
Measurements taken
from reps 1-4 only. No
measurements were
taken from rep 5.
Years:
2008 - starting in June
2009
2010 – data from only 3
reps for G7 and G10
2011
(2012 data unavailable)

Equipment
Table 15. Farm equipment used in field experiment.

80

Table 15 (cont’d)

81

Arlington Agricultural Research Station
Material Inputs
Table 16. Material inputs at ARL per reference flow of 1 ha year-1.

82

Table 16 (cont’d)

83

Non-Material Inputs
Table 17. Non-material
material inputs at ARL per reference flow of 1 ha year-1.

84

Material and Energetic Output
Table 18. ARL: Yield refers to grain yield at standard moisture content for corn, soybeans,
canola; dry biomass yield of perennial crops. Per reference flow 1 ha year-1.
-1

-1

System

Yield (Mg ha )

Potential Energy (MJ ha )

Source

Notes

G1

12.19

99522.12

WI field data

corn ethanol energy [1]

corn-rot

11.82

96500.60

corn ethanol energy [1]

soybean-rot 3.92

27207.50

soybean biodiesel energy [2]

canola-rot

1.82

29343.53

canola biodiesel energy [3]

G5

4.62

29473.32

equation [4]

G6

4.35

27765.88

equation [4]

G7

3.82

24362.47

equation [4]

G9

2.79

17807.75

equation [4]

G10

3.37

21526.30

equation [4]

85

Nitrous Oxide Emissions
Table 19. ARL: Nitrous oxide emissions from each cropping system. Source: 4.3, in press.
-1

-1

-1

Cropping System

kg N2O-N ha yr

kg N2O ha yr

G1 Continuous Corn

4.8270

7.5853

corn-rot

5.3187

8.3580

soybean-rot

2.6463

4.1585

canola-rot

2.6782

4.2086

rotational system

3.7238

5.8517

G5 Switchgrass

4.0282

6.3300

G6 Miscanthus

2.1941

3.4479

G7 Native Grass mix

2.5905

4.0708

G9 Old Field

1.8647

2.9302

G10 Prairie Grass mix

2.6182

4.1144

-1

Data Years
Years:
2008 - 2 time points only
2009 - miscanthus didn’t
overwinter; may have
cause inaccuracies in G6
data
2010
2011
2012

General Assumptions and GaBi 6.0 Model Assumptions
General Assumptions
Models relied on the intrinsic values included in the Professional + Extension 2002
database within GaBi 6.0 for all other input and output values associated with material and
non-material inputs and outputs that are not listed in above tables .
The switchgrass biomass conversion equation [4] was used for the biomass to energy
conversions for G7 Native Grass Mix, G9 Old Field, and G10 Prairie Grass Mix.
Every material and/or non-material input was applied equally across entire plot area in
time.

86

Every material and/or non-material output was harvested or released equally across
entire plot area in time.
Gases, sequestered and/or emitted, were sequestered or emitted equally across entire
plot area in time.
Unit processes and aggregated processes (LCI results) from the United States were used
whenever available in the Professional + Extension 2002 database within GaBi 6.0. When United
States derived processes were unavailable, models were built with processes derived from
Canada due to the close geographical proximity and relatively similar infrastructure between
countries. When United States and Canada derived processes were unavailable in the
Professional + Extension 2002 database within GaBi 6.0, European, particularly German, derived
processes were used in the cropping system models.
Equipment was not documented for Arlington Agriculture Research Station location.
Equipment was assumed to be the same at both locations, therefore Kellogg Biological Station
equipment was used for this missing information.

GaBi Model Unit Process Assumptions
Miscanthus
No appropriate model was present in the Professional + Extension 2002 GaBi 6.0
database or could be created to approximate miscanthus rhizomes. Rhizomes were not
included within the model due to these limitations.
No planting information was included for miscanthus planting at either location.
Miscanthus was hand planted. Hand planting was not replicable in the GaBi model.

87

Corn and Corn in Rotation (corn-rot)
Biomass remaining in field after grain harvest (stover) was not modeled due to the
limitations imposed by the software.
Nitrous Oxide Flux Data
There was incomplete nitrous oxide gas flux data for 2008 at both sites:
In 2008, the AARS fluxes are calculated from 2 (rather than the now-standard 4) time
points. The available data was used in the models with no modifications.
In 2008, KBS did not start monitoring until June instead of at the beginning of the
growing season. The available data was used in the models with no modifications.
In 2009, the majority of the miscanthus at Arlington Agricultural Research Station didn’t
overwinter so flux data may not be reliable for that location year. The available data was used
in the models with no modifications.
In 2012, no data was available for KBS. The four-year average (2008, 2009, 2010, 2011)
was used in place of a five-year average value in the models with no modifications.

Energy Conversions
Corn Grain Ethanol
corn grain yield in Mg per hectare × covert to kg × litres per kg of corn grain* × density of ethanol
× energy per liter of ethanol × convert to MJ = energy per hectare
corn grain yield

‫ ݃ܯ‬1000 ݇݃
1 ݈݅‫݁ݎݐ‬
0.789 ݇݃ 26,900 ݇‫ܬ‬
1 ‫ܬܯ‬
‫ܬܯ‬
ൈ
ൈ
ൈ
ൈ
ൈ
ൌ
݄ܽ
1 ‫݃ܯ‬
2.6 ݇݃ ݃‫݊݅ܽݎ‬
1 ݈݅‫݁ݎݐ‬
1 ݇݃
1000 ݇‫݄ܽ ܬ‬
[1]

*conversion factor: Pimental, 2001.
88

Soybean Biodiesel
soybean yield in Mg per hectare × convert to kg × convert to pounds × pounds per bushel of soybeans
× gallons of biodiesel per bushel* × energy per gallon of biodiesel × convert to MJ = energy per hectare
soybean yield

1 ܾ‫ݑ‬
1.5 ݈݈݃ܽ‫ ݏ݊݋‬119500 ܾ‫ݑݐ‬
‫ ݃ܯ‬1000 ݇݃ 2.2 ݈ܾ‫ݏ‬
ൈ
ൈ
ൈ
ൈ
ൈ
1 ‫݃ܯ‬
1 ݇݃
60 ݈ܾ‫ݏ‬
1 ܾ‫ݑ‬
1 ݈݈݃ܽ‫݊݋‬
݄ܽ
ൈ

1 ‫ܬܯ‬
‫ܬܯ‬
ൌ
947.817120313 ܾ‫݄ܽ ݑݐ‬
[2]

*conversion factor: Pradhan et al. 2009.

Canola Biodiesel
canola yield in Mg per hectare ൈ convert to kg ൈ conver to pounds
ൈ pounds per bushel of canola ൈ gallons of biodiesel per bushel*
ൈ energy per gallon of biodiesel ൈ convert to MJ ൌ energy per hectare
canola yield

‫ ݃ܯ‬1000 ݇݃ 2.2 ݈ܾ‫ݏ‬
1 ܾ‫ݑ‬
2.9 ݈݈݃ܽ‫ ݏ݊݋‬119500 ܾ‫ݑݐ‬
ൈ
ൈ
ൈ
ൈ
ൈ
݄ܽ
1 ‫݃ܯ‬
1 ݇݃
50 ݈ܾ‫ݏ‬
1 ܾ‫ݑ‬
1 ݈݈݃ܽ‫݊݋‬
ൈ

1 ‫ܬܯ‬
‫ܬܯ‬
ൌ
947.817120313 ܾ‫݄ܽ ݑݐ‬
[3]

*conversion factor: McIntosh et al. 1983.

89

Switchgrass Cellulosic Ethanol
biomass yield in Mg per hectare × convert to tons × gallons per ton of switchgrass*
× convert to litres × convert to kg
× energy per litre of ethanol × convert to MJ = energy per hectare
biomass yield

1 ‫݊݋ݐ‬
72 ݈݈݃ܽ‫ ݏ݊݋‬3.78541 ݈݅‫ ݏ݁ݎݐ‬0.789 ݇݃ 26900 ݇‫ܬ‬
‫݃ܯ‬
ൈ
ൈ
ൈ
ൈ
ൈ
1 ‫݊݋ݐ‬
1 ݈݈݃ܽ‫݊݋‬
1 ݈݅‫݁ݎݐ‬
1 ݇݃
݄ܽ 0.90718 ‫݃ܯ‬
ൈ

1 ‫ܬܯ‬
‫ܬܯ‬
ൌ
1000 ݇‫݄ܽ ܬ‬
[4]

*conversion factor: from Morrow et al. 2006.

Corn Stover Cellulosic Ethanol
biomass yield in Mg per hectare × convert to tons × gallons per ton of stover*
× convert to litres × convert to kg
× energy per litre of ethanol × convert to MJ = energy per hectare
biomass yield

‫݃ܯ‬
1 ‫݊݋ݐ‬
80 ݈݈݃ܽ‫ ݏ݊݋‬3.78541 ݈݅‫ ݏ݁ݎݐ‬0.789 ݇݃ 26900 ݇‫ܬ‬
ൈ
ൈ
ൈ
ൈ
ൈ
݄ܽ 0.90718 ‫݃ܯ‬
1 ‫݊݋ݐ‬
1 ݈݈݃ܽ‫݊݋‬
1 ݈݅‫݁ݎݐ‬
1 ݇݃
ൈ

1 ‫ܬܯ‬
‫ܬܯ‬
ൌ
1000 ݇‫݄ܽ ܬ‬
[5]

*conversion factor from: Sheaffer et al. 2010.

90

REFERENCES

91

REFERENCES
Bare, Jane C. “Developing a consistent decision-making framework by using the US EPA’s TRACI.”
Cincinnati, Ohio, USA: US Environmental Protection Agency (2002).
Bare, J.C., Norris, G.A., Pennington, D.W., and T. McKone. “TRACI: The Tool for the Reduction and
Assessment of Chemical and Other Environmental Impacts”. Journal of Industrial Ecology
(2003). 6:3-4.
Bessou, Cécile, et al. “Using a crop model to account for the effects of local factors on the LCA of sugar
beet ethanol in Picardy region, France.” The International Journal of Life Cycle Assessment 18.1
(2013): 24-36.
Brentrup, P., and C. Palliere. 2008. “GHG emissions and energy efficiency in European nitrogen fertiliser
production and use. Proc: International Fertiliser Society, December 11, York, UK.
Caffrey, Kevin R., and Matthew W. Veal. “Conducting an Agricultural Life Cycle Assessment: Challenges
and Perspectives.” The Scientific World Journal 2013 (2013).
Carson, Walter P., and Gary W. Barrett. “Succession in old-field plant communities: effects of contrasting
types of nutrient enrichment.” Ecology 69.4 (1988): 984-994.
Congressional Research Service. 2013. “Renewable Fuels Standard: Overview and Issues.” March 14.
www.fas.org/sgp/crs/misc/R40155.pdf (accessed April 11, 2014).
Drapcho, C.M.; Nghiem, N.P.; Walker, T.H. Biofuels Engineering Process Technology; McGraw-Hill: New
York, NY, USA, 2008.
Gelfand, I. and G. P. Robertson. 2014. Mitigation of Greenhouse Gases in Agricultural Ecosystems. Pages
forthcoming. S. K. Hamilton, J. E. Doll, and G. P. Robertson, editors. The ecology of agricultural
ecosystems: long-term research on the path to sustainability. Oxford University Press, New
York, New York, USA.
Great Lakes Bioenergy Research Center. 2014. “Experimental Layouts.” data.sustainability.glbrc.org
(accessed April 11, 2014).
Great Lakes Bioenergy Research Center. 2014. “GLRBC - About”. www.glbrc.org/about (accessed April
11, 2014).
Hedtcke, J.L., G.R. Sanford, K.E. Hadley, and K.D. Thelen. 2014. Maximizing land use during switchgrass
establishment in the North Central United States. Agron J. 106:596-604.
Huang, Ying, Yaying Li, and Huaiying Yao. “Nitrate enhances N2O emission more than ammonium in a
highly acidic soil.” Journal of Soils and Sediments 14.1 (2014): 146-154.
Hutchinson, G.L., Mosier, A.R., 1981. Improved Soil Cover Method for Field Measurement of Nitrous
Oxide Fluxes. Soil Sci. Soc. Am. J. 45, 311. doi:10.2136/sssaj1981.03615995004500020017x

92

Intergovernmental Panel on Climate Change (IPCC). 2007. “2.10.2 Direct Global Warming Potentials.”
http://www.ipcc.ch/publications_and_data/ar4/wg1/en/ch2s2-10-2.html (accessed April 16,
2014).
Kaltschmitt, M., Reinhardt, G.A., and T. Stelzer. Life cycle analysis of biofuels under different
environmental aspects. Biomass and Bioenergy (1997); 12:121-34.
Kendall, A. and J. Yuan. “Comparing life cycle assessments of different biofuel options.” Current Opinion
in Chemical Biology (2013); 17:439-443.
Lemke, Reynald, and Rich Farrell. “Nitrous Oxide Emissions and Prairie Agriculture.” Prairie Soils and
Crops 1 (2008): 11-15.
Livingston, G., Hutchinson, G.P., 1994. Enclosure-based measurement of trace gas exchange:
applications and sources of error, in: Matson, P.A., Harriss, R.C. (Eds.), Biogenic Trace Gases:
Measuring Emissions from Soil and Water. Blackwell Scientific, London, pp. 14 – 51.
Lohbeck, Madelon, et al. “Changing drivers of species dominance during tropical forest succession.”
Functional Ecology (2014).
McIntosh, Christopher S., Stephen M. Smith, and Russell V. Withers. “Energy balance of on-farm
vegetable oil production and extraction in selected areas of Idaho and Washington.” Research
Bulletin, Idaho Agricultural Experiment Station 129 (1983).
Menten et al. “A review of LCA greenhouse gas emissions results for advanced biofuels: The use of
meta-regression analysis.” Renewable and Sustainable Energy Reviews 2013
Michigan State University Enviro-weather. 2014. “Enviro-weather.” www.agweather.geo.msu.edu
/mawn/sation.asp?id=kbs (accessed April 8, 2014).
Morrow, William R., W. Michael Griffin, and H. Scott Matthews. “Modeling switchgrass derived cellulosic
ethanol distribution in the United States.” Environmental Science & Technology 40.9 (2006):
2877-2886.
ISO, EN. “14040: 2006.” Environmental management-Life cycle assessment-Principles and framework.
European Committee for Standardization (2006).
ISO, EN. “14044: 2006.” Environmental management-Life cycle assessment-Requirements and
guidelines. European Committee for Standardization (2006).
National Climatic Data Center. “1981-2012 Normals Data Access.” http://www.ncdc.noaa.gov/landbased-station-data/climate-normals/1981-2010-normals-data (accessed April 18, 2014).
National Weather Service. 2014. “National Weather Service Forecast Office.” www.nws.noaa.gov/
climate (accessed April 8, 2014).
Oates et al. “Nitrous oxide emissions from model biofuel cropping systems in the establishment phase.”
(working title 2014).

93

Parkin, T.B., and Rodney T. Venterea. “USDA-ARS GRACEnet Project Protocols Chapter 3. Chamber-Based
Trace Gas Flux Measurements 4.” Sampling Protocols. USDA-ARS, Fort Collins, CO (2010): 3-1.
Pimentel, David. “Biomass utilization, limits of.” Encyclopedia of physical science and technology 2
(2001): 159-171.
Pradhan, A., Shrestha, D.S., McAloon, A., Yee, W., Hass, M., Duffield, J.A., and H. Shapouri “Energy lifecycle assessment of soybean biodiesel.” United States, Department of Agriculture, Agricultural
Economic Report 845 (2009).
Renewable Fuels Association. 2014. “Industry Resources: Co-products.” www.ethanolrfa.org/pages/
industry-resources-coproducts (accessed April 14, 2014).
Robertson, G.P., E.A. Paul, and R. R. Harwood. “Greenhouse gases in intensive agriculture: contributions
of individual gases to the radiative forcing of the atmosphere.” Science 289.5486 (2000): 19221925.
Sheaffer, C., J. Lamb, and C. Rosen. “Corn Stover: Ethanol Production and Nutrient Uptake.” December
2010. www.mncorn.org/.../mncorn.../B6-CornStoverEthanolProducƟon_2.pdf. (accessed April
29, 2014).
Skowroñska, Monika, and Tadeusz Filipek. “Life cycle assessment of fertilizers: a review.” Int Agrophys
28 (2014): 101-110. doi: 10.2478/intag-2013-0032.
Smith, K.A., Dobbie, K.E., 2001. The impact of sampling frequency and sampling times on chamber-based
measurements of N2O emissions from fertilized soils. Glob. Chang. Biol. 7, 933–945.
doi:10.1046/j.1354-1013.2001.00450.x.
Stein, D. 2012. 2013 Custom Machine and Work Rate Estimates. Michigan State University Extension.
Toma, Yo, et al. “Nitrous oxide emission derived from soil organic matter decomposition from tropical
agricultural peat soil in central Kalimantan, Indonesia.” Soil science and plant nutrition 57.3
(2011): 436-451.
U.S. Department of Energy. 2011. U.S. Billion-ton update: Biomass supply for a bioenergy and
bioproducts industry. R.D. Perlack and B.J. Stokes, editors, ORLN/TM-2011/224. Oak Ridge Natl.
Lab., Oak Ridge, TN. 227p. http://info.ornl.gov/sites/publications/files/Pub31057.pdf (accessed
28 Oct. 2013).
Wright, L. “Historical perspective of how and why switchgrass was selected as a “model” high-potential
energy crop”. 2007.

94

CHAPTER 2
MAXIMIZING LAND USE DURING SWITCHGRASS ESTABLISHMENT
IN THE NORTH CENTRAL UNITED STATES

Reprinted by Permission, ASA, CSSA, SSSA. A reformatted version of this chapter was
published in the Agronomy Journal as: Hedtcke, J.L., G.R. Sanford, K.E. Hadley, and K.D. Thelen.
2014. Maximizing land use during switchgrass establishment in the North Central United States.
Agron J. 106:596-604.
Abstract: Switchgrass (Panicum virgatum L.) grown under three differing establishment
methods was evaluated for yield, quality, and potential ethanol production across seven
environments in Wisconsin and Michigan. The three establishment methods included: (i) June
seeding into fallow ground; (ii) June seeding following winter rye (Secale cereale L.) forage; and
(iii) August seeding following winter wheat (Triticum aestivum L.). Planting switchgrass in June
was successful following the rye forage crop, but planting in August after wheat resulted in
complete stand loss. While harvested biomass in a two-cut system was equal to or greater than
that realized in a one-cut system (6.9 vs. 6.8 Mg ha–1), biomass quality for ethanol production
was highest following a killing frost (226 vs. 224 g ethanol kg–1 biomass). The higher overall
biomass production in the two-cut system generally compensated for this difference however,
with ethanol yields similar between the two systems (1950 vs. 1970 L ethanol ha–1 for the twoand one-cut system, respectively). In addition to ethanol production, we found that forage
nutritive value in the first cut of the two-cut system was of sufficient quality to satisfy the
dietary needs of several classes of livestock. Harvesting established switchgrass for hay with an
early season cutting in a two-cut system provides producers with an alternative forage source
95

while not affecting total seasonal biomass yield if harvested at the appropriate growth stage
and cutting height to leave sufficient photosynthetic material for regrowth.

INTRODUCTION
Ethanol and other biofuels have the potential to replace 30% or more of the gasoline
demand in the United States by 2030 (U.S. Department of Energy, 2011). Presently, the
majority of domestic ethanol comes from corn (Zea mays L.) grain, but the long-term
sustainability of grain ethanol is questionable as it may increase competition between food,
feed, and fuel interests (USDA-ERS, 2013). Switchgrass has been identified as a candidate
bioenergy crop in recent decades by the U.S. Department of Energy, particularly as a secondgeneration feedstock for conversion to liquid fuels (Monti, 2012). Recent advances in cellulosic
conversion technology and the development of large-scale cellulosic ethanol production
facilities have made switchgrass a viable alternative to corn grain ethanol (McLaughlin et al.,
2002). Switchgrass is attractive as a biofuel crop because of its adaptability from Canada to
Texas, its ability to grow on marginal, highly erodible, and drought-prone soils, and its
environmental benefits which include C sequestration and wildlife habitat (McLaughlin and
Walsh, 1998). It can also be grown and harvested using existing forage production equipment,
reducing potential barriers to farmer adoption.
A significant number of producers surveyed in Tennessee and Illinois indicated that as
production technology and technical assistance become increasingly available, and as demand
creates a viable market for biomass, they would be interested in growing switchgrass on their
farms (Sanderson et al., 2012). They may, however, be hesitant to do so if income losses cannot

96

be offset during the slow establishment period of switchgrass. Although switchgrass can be
harvested in as little as 2 years from establishment (Vogel et al., 2002), it may not reach its full
yield potential until the third growing season in the colder climates of the north central United
States. Economically feasible systems must therefore be designed which can carry a producer
through the critical establishment period. The purpose of this study was threefold and sought
to: (i) evaluate two cropping options with small grains (winter rye and winter wheat) from
which farmers could realize economic gains during the establishment phase of switchgrass
while not reducing subsequent switchgrass productivity; (ii) to evaluate the potential tradeoffs
of switchgrass managed as a dedicated bioenergy crop (October harvest) vs. a dual-use feed
and biomass crop (two cuts; June and October); and (iii) to determine the quality of biomass
and theoretical ethanol yield produced under the one-cut and two-cut harvest management
strategies. Our initial three hypotheses were that: (H1) A rye forage crop preceding switchgrass
establishment would depress switchgrass yields compared to that following a fallow phase; (H2)
Planting switchgrass in early June (following fallow or rye) would provide a better stand than
planting in August (following winter wheat); and (H3) Late-season biomass from the two-harvest
management system would be of equal or higher quality than the late-season biomass from the
one-harvest management system in terms of cattle (Bos taurus) feed and theoretical ethanol
yield.

97

MATERIALS AND METHODS
Treatments, Experimental Design, and Site Description
Switchgrass yields were evaluated in both the seeding year and first production year
under three treatments. The three treatments consisted of different crop establishment
methods and included switchgrass planting: (i) in June, fallow since previous fall (control
treatment); (ii) in June, following a winter rye forage crop; and (iii) in August, following wheat
harvested for grain and straw. The experimental design was a randomized complete block, with
four replicates. To experience a wider range of seasonal rainfall and temperature patterns, as
well as different soil types and field management, the trial was initiated seven times at various
locations. All sites had been in field crop rotations for several years before the study (Table 1).
For most sites, the April to August rainfall was above average in 2010 and below average in
2011. In the fall of 2009, three sites in southern Wisconsin (Arlington [ARL], Rio [RIO], and Sugar
Creek [SC]) were established. In the fall of 2010, three more sites nearby or adjacent to the
2009 fields plus an additional site in southwest Michigan (Kellogg Biological Station, KBS) were
added. The combination of site and year were merged into one variable referred to as
“environment”. Soils were assessed for fertility before the start of the trial at each environment
and were all in an “optimum” range for switchgrass production based on University of
Wisconsin-Extension recommendations (Laboski and Peters, 2012). Soil test phosphorus (STP),
soil test potassium (STK), organic matter (OM), and pH status along with other basic site
characteristics are shown in Table 1. All environments were managed with no-till practices
except SC-I and SC-II.

98

Table 20. Site characteristics from seven environments in southern Wisconsin and lower
Michigan.

In addition to having a recent cropping history with good fertility (see Table 1), no major
perennial weed issues existed which created ideal conditions for switchgrass establishment. In
the established year of switchgrass (Year 2), a split plot was imposed over the three previous
crop treatments (fallow, rye, wheat). Subplots consisted of two harvest managements: (i) a
two-cut management system harvested once in June, targeted at early boot stage (R0 stage per
Moore et al., 1991) and once again after the killing frost (S5 growth stage per Moore et al.,
1991) and (ii) a one-cut management system, harvested after the killing frost (S5).

Small Grains Management
Winter rye and winter wheat were planted on the same day at each given environment
according to UW-Madison extension best management practices using a conventional or no-till

99

grain drill. While the winter wheat variety was the same at all environments (Pioneer 25R47),
the rye varieties were chosen based on regional adaptation and differed by state (WI: cultivar
Spooner, MI: cultivar Wheeler). Plot size averaged 9 by 30 m, large enough to accommodate
field-scale planting and harvest equipment. Glyphosate (N-[phosphonomethyl] glycine) was
applied at a rate of 0.95 kg a.e. ha–1 as a burndown treatment before seeding wheat and rye.
Across the environments, a subsequent post emergence application of a broadleaf herbicide
such as Harmony Extra XP (thifensulfuron-methyl [93 g a.i. ha–1] + tribenuron-methyl [47 g a.i.
ha–1]) or 2,4-D (2,4-Dichlorophenoxyacetic acid [1.07 kg a.e. ha–1]) was applied at the jointing
leaf stage of wheat. Herbicide was not necessary in the winter rye treatment. The exception to
herbicide use was the SC environments where no herbicide was used throughout the trial. The
fallow treatment was not sprayed or tilled during the small grain phase. At most environments,
N was applied to the wheat and rye treatments in early spring at 67 kg N ha–1 as urea. Being
part of SC farm’s overall management practices, manure was applied evenly across the field
before planting the small grains at SC at an equivalent available N rate to the fertilizer used at
other sites. Minimal effect of the residual manure nutrients was expected nor observed.
Winter rye forage was harvested at boot stage in mid-to-late May across environments
to provide maximum total digestible nutrients (i.e., best combination of yield and nutrition) for
livestock feed (McCormick et al., 2006). A center strip from each of the rye plots was cut with a
self-propelled sickle bar mower set at a cutting height of 7.5 cm. After weighing the entire fresh
sample on a pad scale, a grab sample of the forage was collected, weighed fresh, and dried at
60°C until it reached a constant weight to determine moisture content. Biomass is reported on
a dry matter (DM) basis. Winter wheat was harvested for grain using field-scale combines in

100

late July or early August. Wheat straw was removed from the field within 3 days of grain
harvest at most environments other than at RIO where straw yields were minimal and
therefore chopped and evenly redistributed back onto the field to maintain OM and STK.

Switchgrass Management
After the cereal crops were harvested, a burndown herbicide treatment of glyphosate
(0.95 kg a.e. ha–1) was applied to control weeds in all treatments at all environments except SC
where tillage was used for weed control. A second burndown application was occasionally
necessary just before switchgrass seedlings had emerged in the rye or fallow plots. An upland
variety of switchgrass, cultivar Cave-in-Rock was seeded at a depth of 1.3 cm in 19-cm rows in
mid-to-late June with a no-till drill at all sites with the exception of SC where a conventional
grain drill was used. At all of the Wisconsin environments, the rate on the drill was set at 22 kg
ha–1. This rate was set approximately twice as high as the recommended rate to compensate for
poor germination (12%) and a high percentage of hard seed (78%) in the seed lot purchased. A
different seed lot, with higher germination (∼80%), was planted at KBS and the grain drill was
therefore set at 9 kg seed ha–1.
Though the soil was not tested for residual nitrate N, it was decided not to apply any N
during the switchgrass seeding year to discourage weeds and to limit unnecessary economic
inputs (Mitchell et al., 2008; Renz et al., 2009; Sanderson and Reed, 2000). When the
switchgrass reached the three to four leaf stage, that is, V3 to V4 per Moore et al. (1991) (late
July/early August), 2,4-D (0.53 kg a.e. ha–1) or a combination of Quinclorac (3,7-dichloro-8quinolinacarboxylic acid [0.42 kg a.i. ha–1]) and atrazine (2-chloro-4-[ethylamino]-6-

101

[isopropylamino]-s-triazene [0.12 kg a.i. ha–1]) was used on the rye and fallow plots to control
broadleaf weeds. The exception was at the SC environment where weeds were managed by
clipping with a rotary mower. In the first production year of switchgrass, before emergence, a
spring burndown treatment of glyphosate was applied across environments (except SC and
KBS) to reduce weed pressure from winter annuals and cool-season grasses. A follow-up
broadleaf herbicide application of 2,4-D at a rate of 0.53 to 1.07 kg a.e. ha–1 was made in late
May or early across environments (except SC and KBS). No N was applied to the switchgrass in
the production year other than at KBS where 79 kg N ha–1 was applied in June.

Forage Sampling
In the seeding year, three subsamples of switchgrass were collected from each
experimental unit (i.e., plot) during peak biomass (late September) with hand shears and 0.5-m2
quadrat. Weeds were sorted and separated from switchgrass and the biomass of each was
recorded (no weed data from ARL-II and RIO-II). During the established phase (second year),
three subsamples of switchgrass were collected from each of the 16 plots (four replicates × two
previous crops × two harvest managements) with a 0.25-m2 quadrat and hand shears at a 15cm stubble height. The first cutting of the two-cut harvest treatment was sampled in June when
the switchgrass reached boot stage (about 45–60 cm in height). This harvest strategy was based
on the work of Vogel (2004) that demonstrated maximized seasonal switchgrass production
and persistence under such management in Nebraska. As soon as possible after sampling in
June, field-scale haying equipment removed the biomass from the “two-cut” system plots at a
target cutting height of 15 cm. The second cutting in the two-cut system plots and the one-cut

102

system plots were sampled after the killing frost (–3°C) in October. Samples were collected
from the center of the plots to eliminate shading and border effects. Samples were then dried
at 60°C, in a forced-air oven until constant mass was achieved. Results are reported on DM
basis. Dry samples were ground through a 1-mm screen in a Thomas Wiley hammer mill (Arthur
H. Thomas, Philadelphia, PA) and retained for subsequent polysaccharide analysis. The nutritive
value and mineral composition of whole-plant switchgrass forage was analyzed on all samples
at the University of Wisconsin Soil and Forage Analysis Laboratory (Marshfield, WI) using near
infrared reflectance spectrophotometer (NIRSystem, Inc., Silver Spring, MD, model 6500) to
predict the following quality parameters: crude protein (CP), neutral detergent fiber (NDF), and
neutral detergent fiber digestibility (NDF-D). Whole-plant K was measured by atomic absorption
spectroscopy (AOAC, 2008; Schulte et al., 1987).

Soil Moisture
Volumetric moisture content (VMC, g kg–1) status of the soil was collected at a subset of
environments (n = 4) that ranged most widely in soil type and seasonal moisture. Soil moisture
was measured with a SpectrumTechnologies, Inc. (Aurora, IL) Field Scout 300 time domain
reflectometer (TDR) probe on all treatments at two depths (12 and 20 cm) within 1 week after
seeding (i.e., June and August) in four subsamples per plot. Data from ARL and RIO were used
to compare soil type and effect of treatments on VMC in the two seasons.

Switchgrass Stand Frequency
Switchgrass establishment success was evaluated using the grid method (a grid of 25
cells of 0.15 cm2 each) developed by Vogel and Masters (2001). Frequency counts were taken in
103

the year after seeding to determine if seedlings overwintered and frequencies were above the
40% threshold level, which qualifies as a successful establishment, as determined by others
(Hyder et al., 1971; Schmer et al., 2006). Two measurements were taken from each
experimental unit in mid-July or early August when switchgrass was actively growing, but
before it got too tall to effectively use the grid method.

Theoretical Ethanol Yield
The dried and ground switchgrass samples were sent to the Cell Wall Facility at
Michigan State University for analysis. Acid hydrolysis with a weak acid was used to facilitate
the separation of lignin, hemicellulose, and crystalline cellulose (Foster et al., 2010). The
composition of hemicellulose was determined by an alditol acetate derivation of the soluble
monosaccharides and quantified by gas chromatography-mass spectrometry (GC/MS)
(Albersheim et al., 1967). Residue from the initial acid hydrolysis was washed with an Updegraff
reagent and hydrolyzed with sulfuric acid; a colorimetric assay was used to determine the
crystalline cellulose content of the residue (Updegraff 1969; Selvendran and O’Neill, 1987). The
resulting glucose, xylose, and crystalline cellulose concentrations were used to calculate the
predicted ethanol yield for each sample.
Theoretical ethanol yield was calculated based on the empirically derived hydrolyzable
glucose and xylose levels using Eq. [6] below:
ሼሺሾGlc] + [Cry Cellu] × Glc convሻ ൅ ሺሾXyl] × Xyl conv)ሽ ൈ 51.1%
ൈ metabolic yield = (ethanol mg kg-1 ሻ ሾ6ሿ

104

where [Glc] is the glucose concentration of the biomass following pretreatment and acid
hydrolysis (mg kg–1), [Xyl] is the xylose concentration of the biomass following pretreatment
and acid hydrolysis (mg kg–1), and [Cry] is the crystalline cellulose concentration of the biomass
following pretreatment and acid hydrolysis (mg kg–1). Glc conv is the glucan conversion (%)
following enzymatic hydrolysis of feedstock material and Xyl conv is the xylan conversion (%)
following a separate enzymatic hydrolysis and fermentation (SHF) of feedstock material. Glc
and Xyl conversion values were determined from Jin et al. (2010), for switchgrass (66.5 and
74.7% for Glc and Xyl, respectively). The mass conversion of fermentable sugars to ethanol is
51.1%, and metabolic yield equals the ratio of ethanol to sugars consumed in the fermentation
process divided by 51.1% (Lau and Dale, 2009). The metabolic yield value for switchgrass
(89.7%) was determined using a SHF process and was derived from Jin et al. (2010). The
theoretical ethanol yield was determined by multiplying the result from Eq. [5] with its
corresponding dry matter yield.

Statistical Analysis
Analysis of switchgrass production, VMC, stand frequency, biomass quality, and
theoretical ethanol yield were all conducted using PROC MIXED of SAS 9.1.3. The UNIVARIATE
procedure (SAS 9.1.3) was used to confirm model assumption of normality and variance
equality (SAS Institute, 2006; Littell et al., 2006). Kenward–Roger type degrees of freedom
estimates were used per Gbur et al. (2012). In the mixed model for yield, stand frequency,
biomass quality, and theoretical ethanol yield, both previous crop and harvest management
were considered fixed effects, while environment (site-start combination), block, and their

105

interaction terms were treated as random effects. Treating environment as a random effect
increased the zone of inference for this study to include much of the North Central United
States. Best linear unbiased predictors (BLUP’s) were used to evaluate the random effect of
previous crop and harvest management treatments in individual environments (Littell et al.,
2006). In the case of VMC, due to data from only two locations and 2 years, all terms in the
model, with the exception of block, were treated as fixed. Statistical significance was based on
an α level of 0.05 unless otherwise noted.

RESULTS AND DISCUSSION
Seeding Year Switchgrass
The August seeding of switchgrass failed as a result of winter kill at all seven
environments despite successful germination and late season biomass establishment (∼10 cm
of growth). Stand frequency 1 year after seeding averaged just 5.6% for this treatment.
Competition from volunteer winter wheat, and the lack of time between planting and killing
frost for root establishment likely contributed to the failed seeding. Furthermore, the
morphological development of warm-season grasses in which the coleoptilar node is relatively
short makes them relatively sensitive to the dry soil conditions commonly encountered during
the late summer in the North Central United States (Hoshikawa 1969).
There was no effect of previous crop (fallow vs. rye) on seeding year switchgrass yields
(P = 0.3602). In the successful treatments, switchgrass seeding-year yields ranged dramatically
across environments from <0.11 Mg ha–1 to more than 3.37 Mg ha–1. Poor weed control at SC
combined with a dry summer in 2011 hindered switchgrass growth at that environment;

106

conversely, the consistent rainfall throughout 2010 and the warm sandy soils at RIO combined
to provide ideal growing conditions for the switchgrass planting (Figure 39).

Figure 39. Seeding year switchgrass yields at each of the seven environments. Values
represent the mean of four replicates, and error bars represent 1 standard error of the mean.

Weed control is especially important in a slow-establishing perennial such as
switchgrass. There was no effect of previous crop on weed biomass within the switchgrass plots
(P < 0.1318). Overall, weeds comprised about 30% the total seeding year biomass. As shown in
Figure 39, at the SC-II environment where weeds were managed without herbicides in a very
dry season, second year switchgrass production was depressed compared to the other
environments that used herbicides in the seeding year.

107

Surface Soil Moisture
In 2010, a season with ample rainfall and an above average snowfall during the
2009/2010 winter, VMC at 12-cm depth was adequate (above permanent wilting point) and
similar across all treatments and sampling times for each soil type (Figure 40). Striking
differences were evident between years at RIO when June 2011 had much lower surface soil
moisture than June 2010. The actively growing wheat treatment in the relatively dry spring of
2011 had lower VMC than the other treatments across soil types which may have been an
additional factor in the failed switchgrass establishment. The rye treatment was above the
permanent wilting point of 100 g kg–1 (Figure 40). Cover crops, especially cereal crops, are well
documented for their ability to conserve soil moisture through reduced surface evaporation
(Haramoto and Brainard, 2012; Krueger et al., 2011; Unger and Vigil, 1998). Decomposition of
the rye roots may have also improved infiltration and water holding capacity of the soil,
positively affecting early summer VMC readings as shown by other research (Gregory et al.,
2005; Joyce et al., 2002; Lal et al., 1979; Troeh et al., 2004; Unger and Vigil, 1998). A recent
survey of more than 700 growers in the central United States showed that cover crops were
credited with a 10 to 14% improvement in water conservation and soil conditions compared to
field without cover crops (USDA-SARE, 2013). In this same survey, winter cereal grains were the
most common cover crop cited with more than 70% of respondents having reported using
them.

108

Figure 40. Volumetric water content Θ during the 2010 and 2011 growing seasons at: A 12 cm
in June, B 20 cm in June, C 12 cm in August, and D 20 cm in August. Error bars represent the
95% confidence limit for comparison of Θ within a season and soil depth (e.g., A, B, C, D. Solid
line represents field capacity; dotted line represents permanent wilting point for these soils.

Subsurface Soil Moisture
At the 20-cm depth, VMC had a similar pattern as it did at the 12-cm depth, with the
exception that the rye treatment at RIO was lower than the other treatments in the August

109

sampling period (Figure 40). It is possible that the rapidly growing rye was able to pull moisture
from deeper in the profile in the early spring and make it available at shallower depths
following herbicide termination of the rye in mid-May.
Fouli et al. (2012) studying components of soil water balance found deep rooted crops
such as alfalfa (Medicago sativa L.) to have the lowest subsoil VMC compared to more shallowbased crops such as winter rye and corn in dry summers (i.e., half the normal precipitation).

Established Switchgrass: Frequency
Excluding the failed August seeding from the analysis, switchgrass frequency was not
different between the rye and fallow treatments (73 and 70%, respectively; P < 0.1801).
Frequency was lower at SC-I than at other sites but well above the 40% level. However, due to
tremendous seeding year weed pressure and drought the following year, switchgrass frequency
was only 20% at SC-II in 2011.

Established Switchgrass: Production
Second year yields were lower than those reported in the central and southern United
States (Vogel et al., 2002, Fike et al., 2006) and quite variable, averaging 6.8 Mg ha–1 ± 4.2 (SD)
across environments. Similar results were reported by Madakadze et al. (1999), and
Wullschleger et al. (2010), with mean switchgrass yields of 8.7 ± 4.2 Mg ha–1 for “upland type”
cultivars across the United States and eastern Canada. Of the cultivars evaluated, cultivar Cavein-Rock exhibited exceptional high yield variability. We found no lingering effect of the previous
crop history on second year production overall (P < 0.3836) but analysis of BLUPs did show that
yields were equal or higher in the two-cut system in five of the seven environments (Figure 41).
110

The two environments where the one-cut harvest management resulted in significantly higher
yields was due to the June harvest being accomplished later than targeted. Switchgrass
sensitivity to frequent or intensive defoliation has been reported, but researchers in Iowa and
Missouri found that harvesting it in mid-June during stem elongation was not detrimental to
productivity (George and Oberman 1989; Anderson and Matches, 1983). Furthermore,
management practices should be performed so that harvest is not delayed beyond boot stage
as switchgrass advances quickly from “boot” to more sensitive reproductive growth stages.
Regrowth can also be delayed when harvested past boot stage as photosynthetic material is
removed.

Figure 41. Second year switchgrass yields at each of the seven environments. Values represent
the mean of four replicates, and error bars represent 1 standard error of the mean.

111

Double-Cropping and Dual-Use Cropping Systems
A positive yield response to multiple cut management systems have been reported by
others for upland switchgrass varieties, including trials in Iowa, Nebraska, and throughout the
southern United States (Vogel et al., 2002; Fike et al., 2006). Fike et al. (2005) report a 38%
increase in yield, as well as, higher tiller density for Cave-in-Rock switchgrass under a two-cut
vs. a one-cut system. In the current study we found that the two-cut system performed as well
as or better than the one-cut system in five of the seven environments supporting the finding of
others. We did find a significant harvest × previous crop interaction (P < 0.0288). This was
primarily due to higher yields following fallow than rye in the two-cut system, with no effect of
previous crop in the one-cut system.
In addition to cellulosic biomass there is an additional and important livestock feed
component to the multi-cut forage system. Winter rye yields harvested in May at boot stage
averaged 4.6 Mg ha–1 (ranged from 2.45–5.38 Mg ha–1) with acceptable forage quality for most
classes of livestock (McCormick et al., 2006). First-cut hay forage from the switchgrass
harvested at boot stage (late June) yielded 2.2 to 2.5 Mg ha–1 and provided on average124.5
(4.1), 614.4 (5.7), 604.7 (14.9), and 16.7 (0.4) g kg–1 of crude protein, neutral detergent fiber
(NDF), NDF-digestibility, and K, respectively (numbers in parentheses = 1 standard error). These
parameters are similar to a commonly used forage in dairy systems such as orchardgrass
(Dactylis glomerata L.; Hedtcke et al., 2011) and will satisfy the requirements of many classes of
livestock including dry beef cows, gestating beef heifers, lactating beef cows (producing <14 kg
milk d–1), and ewes (Ovis aries) (Buchanan-Smith et al., 2000; Mitchell et al., 2001). Combining
the total forage accumulation over the course of the present study, using winter rye before

112

switchgrass seeding, and taking a June hay harvest in established switchgrass resulted in total
biomass over 11.2 Mg DM ha–1 an increase of more than 38% compared to systems without rye
forage.
The cost of switchgrass establishment is considerable relative to other forage crops. At
present, switchgrass seed is substantially more expensive than typical forage grass and legume
seed, and the slow establishment of switchgrass means that producers must forego a year of
production before they have a harvestable stand. The federal Biomass Crop Assistance Program
created in the 2012 farm bill covers 75% of establishment costs including seed, land
preparation, and planting (Stubbs, 2011). During this study, total establishment costs were
$120 per Mg of dry matter, three times higher than the establishment costs for a cool-season
grass such as orchardgrass.

Biomass Quality and Theoretical Ethanol Yield
Harvested switchgrass biomass averaged 220 g ethanol kg–1 DM. The previous crop, rye
or fallow, did not have an effect on the biomass quality for ethanol production (P = 0.1261).
However, harvest timing did have a significant effect on biomass quality (P < 0.001). In the ARLI, ARL-II, and KBS-I environments, biomass harvested after the killing frost was of higher quality
for ethanol production than the biomass harvested in mid-to-late June. June-harvested biomass
had high CP and low NDF contents sufficient for livestock feed, but inadequate for use as an
ethanol feedstock. Energy conversion technologies favor biomass with low CP and high NDF
concentrations found in late season harvests (Sanderson et al., 1999). The concentration of
inorganic elements, also undesirable in ethanol production, decrease in switchgrass biomass as

113

it matures (Adler et al., 2006). As the concentration of inorganic elements is reduced,
lignocellulose constitutes a greater proportion of the biomass. During conversion, lignocellulose
is broken down into simple sugars for fermentation. The late harvested biomass contained a
higher concentration of lignocellulose and therefore, sugars. The translocation of nutrients,
such as N, from aboveground biomass can serve the dual purpose of increasing biomass quality
for ethanol conversion, and conserving key soil nutrients (Mulkey et al., 2006).
Across all environments, biomass quality did not differ between the late harvested
biomass in the one-cut management system and the late harvested biomass in the two-cut
management system (P = 0.2204). In addition, harvesting switchgrass as a mid-season forage
did not affect the quality of biomass harvested for use in ethanol production when compared to
the post-frost harvested biomass (Figure 42). Although switchgrass regrowth composition was
not tested in our study, Sanderson et al. (1999) determined that CP and NDF concentrations of
switchgrass regrowth were similar to concentrations observed in switchgrass harvested just
once after a killing frost. It is therefore reasonable to expect a two-cut forage and biomass
system to provide adequate livestock nutrition and quality biomass for ethanol conversion.

114

Figure 42. Biomass quality for ethanol production at seven environments across four reps and
two previous crop (i.e., rye or fallow) treatments. Values represent the mean of four replicates,
and error bars represent one standard error of the mean.

Although biomass quality is an important aspect of ethanol production, potential
ethanol yields were more heavily dependent on the amount of biomass harvested than the
composition of the biomass. Potential ethanol yields mirrored the results of second year
switchgrass yields. Ethanol yields ranged from 136 L ha–1 at SC-II (<1 Mg DM ha–1 in the one-cut
system) to 4225 L ha–1 at ARL-II (>15 Mg DM ha–1 in the two-cut system). Although there was no
significant main effect of treatment (on-cut vs. two-cut) on total ethanol yields, analysis by
environment indicated a high degree of variability among sites with significantly higher ethanol
yield in two of the seven environments for each of the harvest systems (Figure 43). Again these
results indicate the tight correlation between total ethanol yield and biomass yields, as
environments with lower fuel yields in the two cut system were also those with lower total
biomass in the two cut system (ARL-I and RIO-I, Figures 41 and 43).

115

Figure 43. Potential ethanol yield at seven environments across four reps and two previous
crop (i.e., rye or fallow) treatments. Values represent the mean of four replicates, and error
bars represent one standard error of the mean.

CONCLUSION
Planting switchgrass in June was successful following a winter rye forage crop. Planting
after winter wheat in August was not however, with significant stand loss observed. The rye
forage crop did not have a negative effect on switchgrass production in either the seeding year
or in the first established year and seemed to conserve soil moisture in dry years. While
harvested biomass in a two-cut system was equal to or greater than that realized in a one-cut
system, biomass quality for ethanol production was highest following a killing frost. In spite of
this result, the higher yield in the two-cut system compensated for the lower biomass quality in
overall ethanol production. Furthermore forage quality in the first cut of the two-cut system

116

was sufficient to meet the dietary needs of several important classes of livestock. Harvesting
established switchgrass for hay with an early season cutting in a two-cut system provides
producers with an alternative forage source while not affecting total seasonal biomass yield if
harvested at the appropriate growth stage and cutting height to leave the sufficient
photosynthetic material for regrowth.

ACKNOWLEDGEMENTS
This work was funded in part by USDA-ARS Specific Cooperative Agreement 58-3655-8692 with the University of Wisconsin-Madison and by the DOE Great Lakes Bioenergy Research
Center (DOE BER Office of Science DE-FC02-07ER64494) and DOE OBP Office of Energy
Efficiency and Renewable Energy (DA-AC05-76RL01830). Research technician Todd Martin
assisted with the KBS site and John Hall and Nicole Tautges assisted with the experiment at the
Wisconsin sites.

CHAPTER 2 NOTE
Although I am not listed as the lead author or co-author, I contributed a significant
amount to this published paper. During my first year as a graduate student, I contributed:
i. writing parts of various sections, including but not limited to, a large portion of the
introduction
ii. stand frequency and quadrat harvests field work and data for the second year at the KBS
location
iii. entire section on theoretical ethanol yield (data acquisition, data analysis, results section)
117

iv. statistical analysis for the entire study with the exception of soil moisture content
v. biomass quality (as it pertained to ethanol production, not the livestock feed section) analysis
For these reasons, I have included this previously published paper as a complete
chapter in my thesis.

118

APPENDIX

119

APPENDIX

Table 21. License agreement for article reprint.
ACSESS-Alliance of Crop, Soil, and Environmental Science Societies LICENSE
TERMS AND CONDITIONS
Apr 03, 2014
This is a License Agreement between Katherine E Hadley (“You”) and ACSESS-Alliance of Crop, Soil, and
Environmental Science Societies (“ACSESS-Alliance of Crop, Soil, and Environmental Science Societies”) provided by
Copyright Clearance Center (“CCC”). The license consists of your order details, the terms and conditions provided
by ACSESS-Alliance of Crop, Soil, and Environmental Science Societies, and the payment terms and conditions.
All payments must be made in full to CCC. For payment instructions, please see information listed at the
bottom of this form.
License Number

3361531363003

License date

Apr 03, 2014

Licensed content publisher

ACSESS-Alliance of Crop, Soil, and Environmental Science Societies

Licensed content publication

Agronomy Journal

Licensed content title

Maximizing Land Use during Switchgrass Establishment in the
North Central United States

Licensed copyright line

Copyright ©2014 by the American Society of Agronomy, Inc.

Licensed content author

Janet L. Hedtcke, Gregg R. Sanford, Katherine E. Hadley, et al.

Licensed content date

Mar 4, 2014

Volume number

106

Issue number

2

Type of Use

Thesis/Dissertation

Requestor type

Author of requested content

Format

Print, Electronic

Portion

chapter/article

Rights for

Main product

Duration of use

0 - 5 years

Creation of copies for the disabled

no

With minor editing privileges

no

For distribution to

United States

In the following language(s)

Original language of publication

With incidental promotional use

no

The lifetime unit quantity of new product

0 to 499

120

Table 21. (cont’d)
The requesting person/organization is:

Katherine E Hadley

Order reference number
Title of your thesis / dissertation

Life-Cycle Assessment of the GLBRC Bioenergy Cropping System /
Maximizing Land Use for Bioenergy Feedstock Production

Expected completion date

May 2014

Estimated size (number of pages)

150

121

REFERENCES

122

REFERENCES
Albersheim, P., D.J. Nevins, P.D. English, and A. Karr. 1967. A method for the analysis of sugars in plant
cell wall polysaccharides by gasliquid chromatography. Carbohydr. Res. 5:340–345.
doi:10.1016/S0008-6215(00)80510-8
Adler, P.R., M.A. Sanderson, A.A. Boateng, P.J. Weimer, and H.G. Jung. 2006. Biomass yield and biofuel
quality of switchgrass harvested in fall or spring. Agron. J. 98:1518–1525.
doi:10.2134/agronj2005.0351
Anderson, B., and A.G. Matches. 1983. Forage yield, quality, and persistence of switchgrass and
caucasian bluestem. Agron. J. 75:119–124. doi:10.2134/agronj1983.00021962007500010030x
AOAC. 2008. Method 985.01. Metals and other elements in plants. AOAC, Gaithersburg, MD.
http://eoma.aoac.org/methods/info.asp?ID=31360 (accessed 20 Aug 2013).
Buchanan-Smith, J., L.L. Berger, C. Ferrell, D.G. Fox, M. Galyean, D.P. Hutcheson, et al. 2000. Nutritional
requirements of beef cattle. 7th revised ed. Natl. Academy of Sci., Washington, DC.
Fike, J.H., D.J. Parrish, D.D. Wolf, J.A. Balasko, J.T. Green, Jr., M. Rasnake, and J.H. Reynolds. 2006.
Switchgrass production for the Upper Southeastern USA: Influence of cultivar and cutting
frequency on biomass yields. Biomass Bioenergy 30:207–213.
doi:10.1016/j.biombioe.2005.10.008
Foster, C.E., T.M. Martin, and M. Pauly. 2010. Comprehensive compositional analysis of plant cell walls
(lignocellulosic biomass) Part I: Lignin. J. Vis. Exp. 37, e1745.
http://www.jove.com/video/1745/comprehensivecompositional-analysis-plant-cell-wallslignocellulosic (accessed 20 Jan. 2014). doi:10.3791/1745
Fouli, Y., S.W. Duiker, D.D. Fritton, M.H. Hall, J.E. Watson, and D.H. Johnson. 2012. Double cropping
effects on forage yield and the field water balance. Agric. Water Manage. 115:104–117.
doi:10.1016/j.agwat.2012.08.014
Gbur, E.E., W.W. Stroup, K.S. McCarter, S. Durham, L.J. Young, M. Christman et al. 2012. Estimation and
inference in linear mixed models In: E.E. Gbur, W.W. Stroup, K.S. McCarter, S. Durham, L.J.
Young, M. Christman et al., editors, Analysis of generalized linear mixed models in the
agricultural and natural resource sciences. ASA, CSSA, and SSSA, Madison, WI. p. 60–67.
George, J.R., and D. Oberman. 1989. Spring defoliation to improve summer supply and quality of
switchgrass. Agron. J. 81:47–52. doi:10.2134/agronj1989.00021962008100010008x
Gregory, M.M., K.L. Shea, and E.B. Bakko. 2005. Comparing agroecosystems: Effects of cropping and
tillage patterns on soil, water, energy use and productivity. Renewable Agric. Food Sys. 20:81–
90. doi:10.1079/RAF200493
Haramoto, E.R., and D.C. Brainard. 2012. Strip tillage and oat cover crops increase soil moisture and
influence N mineralization patterns in cabbage. HortScience 47:1596–1602.

123

Hedtcke, J., J.L. Posner, W. Coblentz, J. Hall, R. Walgenbach, and J. Davidson. 2011. Orchardgrass ley for
improved manure management in Wisconsin: II. Nutritive value and voluntary intake by dairy
heifers. Agron. J. 103:1106–1114. doi:10.2134/agronj2011.0054
Hoshikawa, K. 1969. Underground organs of the seedlings and the systematics of Gramineae. Bot. Gaz.
130:192–203. doi:10.1086/336490
Hyder, D.N., A.C. Everson, and R.E. Bement. 1971. Seedling morphology and seeding failures with blue
grama. J. Range Manage. 24:287–292. doi:10.2307/3896945
Jin, M.J., M.W. Lau, V. Balan, and B.E. Dale. 2010. Two-step SSCF to convert AFEX-treated switchgrass to
ethanol using commercial enzymes and Saccharomyces cerevisiae 424A(LNH-ST). Bioresour.
Technol. 101:8171–8178. doi:10.1016/j.biortech.2010.06.026
Joyce, B.A., W.W. Wallender, J.P. Mitchell, L.M. Huyck, S.R. Temple, P.N. Brostrom, and T.C. Hsiao. 2002.
Infiltration and soil water storage under winter cover cropping in California’s Sacramento Valley.
Trans. ASAE 45:315–326. doi:10.13031/2013.8526
Krueger, E.S., T.E. Ochsner, P.M. Porter, and J.M. Baker. 2011. Winter rye cover crop management
influences on soil water, soil nitrate, and corn development. Agron. J. 103:316–323.
Laboski, C.A.M., and J.B. Peters. 2012. Nutrient application guidelines for field, vegetable, and fruit crops
in Wisconsin. Publ. A2809. UW-Madison Ext., Madison, WI.
www.soils.wisc.edu/extension/pubs/A2809.pdf (accessed 20 Aug. 2013).
Lal, R., G.F. Wilson, and B.N. Okigbo. 1979. Changes in properties of an Alfisol produced by various crop
covers. Soil Sci. 127:377–382. doi:10.1097/00010694-197906000-00009
Lau, M.W., and B.E. Dale. 2009. Cellulosic ethanol production from AFEXtreated corn stover using
Saccharomyces cerevisiae 424A(LNHST). Proc. Natl. Acad. Sci. USA 106:1368–1373.
doi:10.1073/pnas.0812364106
Littell, R.C., G.A. Milliken, W.W. Stroup, R.D. Wolfinger, and O. Schabenberger. 2006. SAS® for mixed
models. 2nd ed. SAS Inst., Cary, NC.
Madakadze, I.C., K.A. Stewart, P.R. Peterson, B.E. Coulman, and D.L. Smith. 1999. Cutting frequency and
nitrogen fertilization effects on yield and nitrogen concentration of switchgrass in a short
season area. Crop Sci. 39:552–557. doi:10.2135/cropsci1999.0011183X003900020041x
McCormick, J.S., R.M. Sulc, D.J. Baker, and J.F. Beurerlein. 2006. Yield and nutritive value of autumnseeded winter-hardy and winter-sensitive annual forages. Crop Sci. 46:1981–1989.
doi:10.2135/cropsci2006.0140
McLaughlin, S.B., D.G. De La Torre Ugarte, C.T. Garten Jr, L.R. Lynd, M.A. Sanderson, V.R. Tolbert, and
D.D. Wolf. 2002. High-value renewable energy from prairie grasses. Environ. Sci. Technol.
36:2122–2129. doi:10.1021/es010963d

124

McLaughlin, S.B., and M.E. Walsh. 1998. Evaluating the environmental consequences of producing
herbaceous crops for bioenergy. Biomass Bioenergy 14:317–324. doi:10.1016/S09619534(97)10066-6
Mitchell, R.B., J. Fritz, K. Moore, L. Moser, K.P. Vogel, D. Redfearn, and D. Wester. 2001. Predicting
forage quality in switchgrass and big bluestem. Agron. J. 93:118–124.
doi:10.2134/agronj2001.931118x
Mitchell, R.B., K.P. Vogel, and G. Sarath. 2008. Managing and enhancing switchgrass as a bioenergy
feedstock. Biofuels Bioprod. Bioref. 2:530–539. doi:10.1002/bbb.106
Monti, A., editor. 2012. Switchgrass: A valuable biomass crop for energy. Springer-Verlag Publ., London,
UK.
Moore, K.J., L.E. Moser, K.P. Vogel, S.S. Waller, B.E. Johnson, and J.F. Pedersen. 1991. Describing and
quantifying growth stages of perennial forage grasses. Agron. J. 83:1073–1077.
doi:10.2134/agronj1991.00021962008300060027x
Mulkey, V.R., V.N. Owens, and D.K. Lee. 2006. Management of switchgrass-dominated conservation
reserve program lands for biomass production in South Dakota. Crop Sci. 46:712–720.
doi:10.2135/cropsci2005.04-0007
Renz, M., D.J. Undersander, and M.D. Casler. 2009. Establishing and managing switchgrass. UW-Madison
Ext., Madison, WI. www.uwex.edu/ces/forage/pubs/switchgrass.pdf (accessed 20 Aug. 2013).
Sanderson, M.A., J.C. Read, and R.L. Reed. 1999. Harvest management of switchgrass for biomass
feedstock and forage production. Agron. J. 91:5–10.
doi:10.2134/agronj1999.00021962009100010002x
Sanderson, M.A., and R.L. Reed. 2000. Switchgrass growth and development: Water, nitrogen, and plant
density effects. J. Range Manage. 53:221–227. doi:10.2307/4003287
Sanderson, M.A., M. Schmer, V. Owens, P. Keyser, and W. Elbersen. 2012. Crop management of
switchgrass. In: A. Monti, editor, Switchgrass: A valuable biomass crop for energy. SpringerVerlag Publ., London, UK. p. 88–90.
SAS Institute. 2006. 1.3 User’s guide: Statistics. Vol. 9. SAS Inst., Cary, NC.
Schmer, M.R., K.P. Vogel, R.B. Mitchell, L.E. Moser, K.M. Eskridge, and R.K. Perrin. 2006. Establishment
stand thresholds for switchgrass grown as a bioenergy crop. Crop Sci. 46:157–161.
doi:10.2135/cropsci2005.0264
Schulte, E.E., J.B. Peters, and P.R. Hodgson. 1987. Wisconsin procedures for soil testing, plant analysis
and feed and forage analysis. Dep. of Soil Sci. Bull. 6. Univ. of Wisconsin, Madison.
Selvendran, R.R., and M.A. O’Neill. 1987. Isolation and analysis of cell walls from plant material.
Methods Biochem. Anal. 32:25–153. doi:10.1002/9780470110539.ch2

125

Stubbs, M. 2011. Biomass Crop Assistance Program (BCAP): Status and issues. Congressional Res. Serv.
7-5700. R41296. www.fas.org/sgp/crs/misc/R41296.pdf (accessed 11 Nov. 2013).
Troeh, F.R., J.A. Hobbs, and R.L. Donahue. 2004. Soil and water conservation: For productivity and
environmental protection. 4th ed. Pearson Education, Inc., Upper Saddle River, NJ.
Unger, P.W., and M.F. Vigil. 1998. Cover crop effects on soil water relationships. J. Soil Water Conserv.
53:200–207.
Updegraff, D.M. 1969. Semimicro determination of cellulose in biological materials. Anal. Biochem.
32:420–424. doi:10.1016/S0003-2697(69)80009-6
USDA-ERS. 2013. Bioenergy: Findings. USDA-Economic Res. Serv. www.ers.usda.gov/topics/farmeconomy/bioenergy/findings.aspx#.UhTyt-5I3ukF (accessed 21 Aug. 2013).
USDA-SARE. 2013. 2012–2013 cover crop survey. USDA-Sustainable Agric. Res. and Educ.
www.northcentralsare.org/Educational-Resources/From-the-Field/Cover-Crops-Survey-Analysis.
(accessed 20 Aug. 2013).
U.S. Department of Energy. 2011. U.S. Billion-ton update: Biomass supply for a bioenergy and
bioproducts industry. R.D. Perlack and B.J. Stokes, editors, ORLN/TM-2011/224. Oak Ridge Natl.
Lab., Oak Ridge, TN. www1.eere.energy.gov/bioenergy/pdfs/billion_ton_update.pdf (accessed
20 Aug. 2013).
Vogel, K.P. 2004. Switchgrass. In: L.E. Sollenberger, L. Moser, and B. Burson, editors, Warm-season (C4)
grasses. Agron. Monogr. 45. ASA, CSSA, and SSSA, Madison WI.
Vogel, K.P., J.J. Brejda, D.T. Walters, and D.R., Buxton. 2002. Switchgrass biomass production in the
Midwest USA: Harvest and nitrogen management. Agron. J. 94:413–420.
Vogel, K.P., and R.A. Masters. 2001. Frequency grid–A simple tool for measuring grassland
establishment. J. Range Manage. 54:653–655. doi:10.2307/4003666
Wullschleger, S.D., E.B. Davis, M.E. Borsuk, C.A. Gunderson, and L.R. Lynd. 2010. Biomass production in
switchgrass across the United States: Database description and determinants of yield. Crop Sci.
102:1158–1168.

126