(Advances in agronomy 117) donald l sparks (eds ) advances in agronomy 117 academic press, elsevier (2012)

Chapter 1 is a timely and comprehensive review of the agro-nomic and ecological implications of biofuel production that includesimpacts on land use, soil erosion and water quality, nutri

University of California, Davis

Emeritus Advisory Board Members

JOHN S BOYER

University of Delaware

EUGENE J KAMPRATH

North Carolina State, University

Prepared in cooperation with the

American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America Book and Multimedia Publishing Committee

DAVID D BALTENSPERGER, CHAIR

RONALD L PHILLIPSUniversity of Minnesota

LARRY P WILDINGTexas A&M University

CRAIG A ROBERTS MARY C SAVIN APRIL L ULERY

Newark, Delaware, USA

AMSTERDAM • BOSTON • HEIDELBERG • LONDON NEW YORK • OXFORD • PARIS • SAN DIEGO SAN FRANCISCO • SINGAPORE • SYDNEY • TOKYO Academic Press is an imprint of Elsevier

225 Wyman Street, Waltham, MA 02451, USA

32 Jamestown Road, London, NW1 7BY, UK

The Boulevard, Langford Lane, Kidlington, Oxford, OX51GB, UK

Radarweg 29, PO Box 211, 1000 AE Amsterdam, The Netherlands

First edition 2012

Copyright Ó 2012 Elsevier Inc All rights reserved.

No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means electronic, mechanical, photocopying, recording or otherwise without the prior written permission of the publisher

Permissions may be sought directly from Elsevier's Science & Technology Rights ment in Oxford, UK: phone (+44) (0) 1865 843830; fax (+44) (0) 1865 853333; email: permissions@elsevier.com Alternatively you can submit your request online by visiting the Elsevier web site at http://elsevier.com/locate/permissions, and selecting Obtaining permission

Depart-to use Elsevier material

Notice

No responsibility is assumed by the publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should be made

ISBN: 978-0-12-394278-4

ISSN: 0065-2113 (series)

For information on all Academic Press publications

visit our website at store.elsevier.com

Printed and bound in USA

12 13 14 15 10 9 8 7 6 5 4 3 2 1

Numbers in parantheses indicate the pages on which the authors’ contributions begin.

International Rice Research Institute (IRRI), Los Ba~nos, Philippines.

CSIRO Plant Industry, Wembley, Western Australia, Australia.

* Current address: Directorate of Groundnut Research, Junagdh, Gujarat, India

ix

Volume 117 contains six excellent reviews that focus on some of the mostprominent global issues of our time- energy, environment, and foodproduction Chapter 1 is a timely and comprehensive review of the agro-nomic and ecological implications of biofuel production that includesimpacts on land use, soil erosion and water quality, nutrient cycling, andbiodiversity Chapter 2 addresses food safety issues related to mineral andorganic fertilizers with emphasis on the presence of trace metals Chapter 3reviews the mechanisms of nickel uptake and hyperaccumulation by plants,and impacts related to soil remediation Chapter 4 deals with conservationagriculture in the semi-arid tropics, with respect to environmental,economic, and ecological opportunities and challenges Chapter 5 is anoverview of the use of green and brown manures in dryland wheatproduction systems in Mediterranean-type environments Chapter 6 isfocused on the benefits, disadvantages, and strategies for enhancing theproductivity and sustainability of rice-wheat cropping systems in the Indo-Gangetic Plains region of the Indian subcontinent.

I am most grateful to the authors for theirfirst-rate reviews

DONALD L SPARKSNewark, Delaware, USA

xi

Agronomic and Ecological

Implications of Biofuels

Catherine Bonin and Rattan Lal

Contents

3.1 Greenhouse gas emissions from land conversion 9 3.2 Using land enrolled in the Conservation Reserve Program 10 3.3 Biofuels and restoration of degraded lands 12

5.1 Nitrogen and litter/residue management 19 5.2 Nitrogen uptake and biomass removal 20 5.3 Gaseous emissions and volatilization 21

6.4 Diversity at the landscape level 25

7 Biofuels and the Soil Carbon Budget 28

8 Invasive Potential of Bioenergy Crop Species 30

Advances in Agronomy, Volume 117 Ó 2012 Elsevier Inc.

DOI: http://dx.doi.org/10.1016/B978-0-12-394278-4.00001-5

1

Biofuels can be alternative energy sources which simultaneously reduce dence on fossil fuels and mitigate climate change by reducing greenhouse gas (GHG) emissions In the US, over 50 billion liters of ethanol produced in 2010 is mandated to increase to 136 billion liters by 2022 Globally, approximately 33.3 million ha (Mha) of land under production of biofuels in 2008 may increase

depen-to as much as 82 Mha by 2020 Whereas data on the energy ef ficiency and GHG balances for biofuels are available, information on agronomic and ecological consequences of large-scale production of bioenergy crops is sparse Thus, this paper describes the potential effects that bioenergy production may have on ecosystems Conversion of land to biofuel crops may have signi ficant impacts on ecosystem services such soil and water quality, GHG emissions, wildlife habitat, net primary productivity, and biological control, and plant diversity at both the landscape and the regional levels Production of exotic species for feedstock may increase the risk of escape from agriculture and inva- sion into natural ecosystems Several feedstocks, while suitable on the basis of energy and GHG assessments, may have negative ecosystem impacts (i.e., increased N export in the Gulf of Mexico) Bioenergy feedstock may compete with food crops for land, water, and nutrient resources, resulting in higher pri- ces for food as well as potential increases in malnutrition and food insecurity Biofuels can be a sustainable and renewable source of energy, but assessments must include ecological impacts, economic costs, and energetic ef ficiencies.

1 Introduction

Biofuels are widely considered as renewable and sustainable tives to fossil fuels They are touted as an energy production system thatcan have both a positive energy balance while offsetting greenhouse gas(GHG) emissions through carbon (C) sequestration The US set an auspi-cious goal of supplying the equivalent of 30% of the nation's petroleum usefrom biomass, requiring 900 M Mg (1 Mg¼ 106

alterna-g¼ 1 metric ton) to 1billion Mg of feedstock, and predict that the generation of this extrabiomass will be based on increases in crop yields and changes in land use(Perlack et al., 2005; Somerville, 2006) In terms of bioethanol production,corn (Zea mays L.) grain is presently the most common source in the US,with over 50 billion liters produced by 189 plants in 2010 (Fig 1;Renewable Fuels Association, 2011)

However, corn grain will likely be only a temporary feedstock optiondue to land limitations: even if all corn produced in the US were usedfor ethanol production, it would only supply 12% of US gasoline needs(Hill et al., 2006) In addition, the Energy Independence and SecurityAct of 2007 (EISA), which mandates that 136.3 billion liters ofrenewable fuels be produced annually by 2022, has capped contributions

from corn grain at 56.8 billion liters (Sissine, 2007) Therefore, a variety ofother bioenergy feedstocks are also being promoted for the development ofsecond generation biofuels such as perennial grasses, woody species, andagricultural residues (Table 1).

In the decade ending in 2010, bioenergy feedstocks have undergoneintense scrutiny and evaluation to determine the net energy yields andGHG balances through the use of life cycle assessment (LCA) Even thoughthe primary goal of biofuels is to provide energy with a low C footprint,LCAs show that not all biofuels are created equally in terms of energyand GHG fluxes (Adler et al., 2007; Davis et al., 2009) These assessmentssuggest that the varying results in energy and GHG balances may be caused

by differences in species attributes, crop production practices, land usechanges, and conversion technologies (Fargione et al., 2008; Huang et al.,2009)

Corn grain is a feedstock within thefirst generation of biofuels, which arefuels derived from plant sugars, starches and oils (Soetaert and Vandamme,

2009) Other first generation feedstocks include sugarcane (Saccharumofficinarum L.), oil palm (Elaeis guineensis Jacq.), and soybean (Glycine max(L.) Merr.) Primary feedstock options vary by country: the US uses corngrain, Brazil relies on sugarcane, the European Union uses wheat (Triticumaestivum L.) and sugar beet (Beta vulgaris L.), while both China and Canadause wheat and corn grain (Balat and Balat, 2009) Examination of ninefirst generation feedstocks based on ecological, energy, and GHG emissionparameters suggest that tropical species such as sugarcane and oil palm may

be a better option than temperate species such as corn and wheat in terms

Figure 1 United States ethanol production and ethanol plants Data modified from the

Renewable Fuels Association (2011)

Bioenergy species

Yield (Mg ha L1 )

Time to establish (years)

Stand tence (years)

persis-Fertilizer need

Stress tolerance

Ecological concernsSwitchgrass

(Panicum virgatum)

flood invasive potential inwestern US

6.0 (fertile)Miscanthus

curcas)

deforestation

Sources: Achten et al (2007); Achten et al (2008); Angelini et al (2009); Djomo et al (2011); Fillion et al, (2009); Hartemink (2008); Heaton et al (2008); Hoskinson et al (2007); Lewandowski et al (2003); Linderson et al (2007); Openshaw (2000); Papong et al (2010); Parrish and Fike (2005); Tilman et al (2006); USDA-NASS (2011); Whan et al (1976).

of sustainability (de Vries et al., 2010) Corn, while relatively high-yieldingand with a high ethanol productivity, also requires large amounts in inputs(i.e., fertilizers) which reduce its net energy ratio (NER, energyoutputs:energy inputs) to about 1.7 (Liska et al., 2009) Most temperateannual biofuel feedstock (i.e., cereal grains, sugar beet) have NERs thatrange between 1 and 4 (Venturi and Venturi, 2003) In contrast,sugarcane, which can yield over 80 Mg ha1, has an NER of 9.2(Hartemink, 2008) Although sugarcane's energy production indicates it to

be a desirable biofuel feedstock, as with other first generation choices, itsuse as a fuel source directly affects food availability and brings it intocompetition with other food crops for land and other resources (Pimenteland Patzek, 2008)

Second generation feedstocks are suggested as a solution to the problemscaused byfirst generation options and as a way to reduce competition withfood crops for limited resources Second generation, or lignocellulosic feed-stocks, include perennial grasses, trees, and agricultural and forestry residues

or wastes Because the entire plant can be used instead of just the grains,sugars, or fats, energy yields per hectare can be much larger In addition,grass and tree feedstocks may be grown on marginal lands with fewerinputs, which can reduce GHG emissions and energy requirements(Debolt et al., 2009) Switchgrass (Panicum virgatum L.) may have anNER as large as 13.1 and can reduce GHG emissions by 95% whencompared to gasoline, while miscanthus (Miscanthus x giganteus) canproduce 2.6 times more ethanol per hectare than corn grain can and has

a potential NER of 12e66 (Heaton et al., 2008; Schmer et al., 2008;Venturi and Venturi, 2003) Tree species such as willows (Salix spp.) andpoplars (Populus spp.) can produce 11.5 Mg ha1 or more of biomasseach year and may have NERs of over 20 when burned for electricity(Djomo et al., 2011) Residues and waste products, as byproducts ofagricultural and anthropogenic activities, would not require anyadditional lands or resources for production The NER for corn stover ofapproximately 2.2 is lower than that of some other second-generationfeedstocks; however, it is important to note that the corn grain may beused for other purposes (Luo et al., 2009) Nonetheless, excessive residueremoval may reduce agronomic yield and degrade soil quality byaccelerating soil erosion and depleting soil C levels (Blanco-Canqui et al.,2006; Lindstrom, 1986) Thus, residue removal rates must be balancedwith agronomic and environmental requirements Other second-generation feedstocks include jatropha ( Jatropha curcas L.), which is

a drought resistant tree that produces non-edible oils from its seeds thatcould reclaim eroded or degraded land (Openshaw, 2000) It can behigh-yielding under the appropriate environmental conditions, but morework needs to be done to determine the potential of this plant(Trabucco et al., 2010) A third generation of biofuels, microalgae and

cyanobacteria, are being developed for the production of biodiesel (Sayre,

2010) Microalgae are very productive and would be land efficient,although the input costs for water and energy could be high, reducingthe NER to close to or less than one (Batan et al., 2010; Clarens et al.,2011)

While much research has focused on the energy and GHG aspects ofbiofuels, other questions on the sustainability of biofuel production in terms

of ecosystem functioning and services are less commonly addressed Theremay be significant repercussions to many ecosystem properties and services

if there is a large-scale land-use change to bioenergy crop feedstock tion However, the ecological and agronomic implications of large-scalebioenergy cropping systems are not fully understood Both feedstockspecies and management practices will likely affect ecosystem propertiesdifferently Future growing demands will increase costs and reduce theavailability of vital resources such as arable land, water, nutrients, and pesti-cide inputs If bioenergy systems are to be sustainable alternatives to fossilfuels, there must be an equilibrium between energy yields and the impacts

produc-on soil quality, wildlife habitat, nutrient cycling, and water cycling (Clarens

et al., 2010; Lardon et al., 2009)

Although all three generations of biofuels have advantages and vantages in terms of energy balances, GHG emissions, net primary produc-tivity (NPP), and environmental quality, this paper focuses primarily onfirstand second-generation feedstocks The principal objective of this paper is todiscuss the potential consequences that biofuels have on agricultural andecological processes and services, and identify gaps in scientific knowledgethat must be addressed in order to understand the full ecological implica-tions of biofuels Readers are referred to other reviews that discuss specificissues related to energy and GHG balances for biofuels (Adler et al., 2007;Davis et al., 2009; Hill et al., 2006)

disad-2 Ecosystem Functions and Services

Ecosystem functions and services are related, but different.Ecosystem functions are described as the processes within an ecosystem,while ecosystem services are the benefits that humans receive fromecosystem functions (Costanza et al., 1997) A single ecosystem functionmay provide more than one service: for example, the function “soilretention” may both maintain farmland and also prevent soil erosion(de Groot et al., 2002) Examples of ecosystem services include waterand climate regulation, soil formation, nutrient cycling, and foodproduction (Table 2) These ecosystem services have a global value of

an estimated $33 trillion yr1 (Costanza et al., 1997), but many of these

ecosystem services are at risk Fifteen of the 24 ecosystem servicesevaluated by the Millennium Ecosystem Assessment (2005) are beingdegraded or are in decline Human modification of lands throughfarming and urbanization may affect food production, water and air

Table 2 Ecosystem services and related processes, modified from the Millennium Ecosystem Assessment (2005) and de Groot et al (2002)

organic matterPrimary production Transfer of solar energy into plant and

animal biomass

discharges

elements as well as land cover

discharges

through biotic interactions

biotic interactions

animal biomass

Raw materials:

wood,fiber, etc Transfer of solar energy into plant andanimal biomass

animal biomass

Biochemicals,pharmaceuticals

Variation and uses for medicinal plants

quality, climate control, and the availability of natural resources (Foley

et al., 2005; Hooper et al., 2005; Tilman et al., 2001) In addition,decreasing biodiversity may further impact ecosystem service declines(Loreau et al., 2001)

3 Land Use Change

Humans have been altering the appearance of the landscape for sands of years and bioenergy cropping systems are likely to accelerate thechange (Meyer and Turner, 1992; Vitousek et al., 1997) As human popu-lations have increased, so has alteration of the landscape (Fig 2) In 1700,only 5% of land was under urban settlements, with nearly 95% of landexisting in natural or semi-natural states, but by 2000 over one-half ofglobal land was dominated by humans, to form a new biome, the

thou-“anthrome,” one dominated by anthropogenic activities (Ellis et al.,

2010) Between 1850 and 1990, the global agricultural land areaquadrupled to 1360 Mha, with land use change (LUC) for agricultural

Natural Vegetation 1700 1850 1990 Forest/ShrublandGrassland

Desert Agriculture

activities releasing approximately 105 Pg C (petagram¼ 1015

g¼ 1 GT) byvegetation and soil C losses (Houghton, 1999)

The amount of land in the US planted to corn will remain at high levels of 37 Mha due to high demand for ethanol and corn exportsand approximately 35e40% of the US corn crop will be used for ethanolproduction (USDA, 2010) Increases in ethanol production may come fromeither increases in feedstock yields or increases in land under feedstockproduction An estimated 30% of the predicted increase in future cornproduction may be due to higher yields, which would be a preferredmethod of ethanol production increases, as it would not cause increasedcompetition for land resources (Keeney and Hertel, 2009) However,yield gains can only increase production to a certain point, and oncereached, future increases in feedstock production may result in LUC In

record-2008, 33.3 Mha were under biofuel production worldwide, but by 2020the land area is expected to increase to as much as 82 Mha (Fargione

et al., 2010) In the US, cropland, pasture, and idle land area is expected

to decrease, while area for perennial crops will increase by as much as20.2 Mha (Perlack et al., 2005) Conversion of a large area of land tofeedstock production will have repercussions on GHG emissions as well

as a variety of ecosystem processes such as C and N cycling, waterquality, and wildlife habitat

3.1 Greenhouse gas emissions from land conversion

The impact that biofuels have on GHG emissions has been widelyaddressed by several papers (Adler et al., 2007; Hoefnagels et al., 2010; Whi-taker et al., 2010) and is briefly outlined here in terms of how LUC affectsemissions Production of bioenergy feedstocks may affect LUC directly orindirectly Direct LUC emissions stem from the conversion of land to feed-stock plantations Indirect emissions occur as a result of additional landconversion that occurs from the land uses or ecosystems displaced by thebioenergy plantation (Melillo et al., 2009; Searchinger et al., 2008).Increased bioethanol demand in the US will affect the allocation of USland area under other crops and ecosystems, but will also indirectly causechanges in land uses around the globe (Keeney and Hertel, 2009) DirectLUC impacts on GHG emissions and their corresponding C debt may

be calculated by estimating loss of soil C stock due to land conversion(Fargione et al., 2008), but calculating indirect effects is morechallenging Indirect LUC is a large source of uncertainty when assessingthe impacts of bioenergy crops, and due to this uncertainty, currentpredictions may be underestimating indirect emissions by as much as140% (Plevin et al., 2010)

Direct LUC can have significant impacts on GHG emissions, as landconversion results in a large initial release of GHG (i.e., a“C debt”) that

may take decades or centuries to recover (Fargione et al., 2008) Withouttaking LUC into account, some corn and cellulosic ethanol productionsystems show predicted 20% and 70% reductions in GHG compared togasoline, but after factoring in LUC, these two ethanol systems mayactually release 93% and 50% more GHGs (Searchinger et al., 2008), butthese calculations have been debated (Mathews and Tan, 2009).Although biofuel feedstock may sequester C and reduce emissions, theinitial costs of land conversion may make these systems more costly in Cterms for a significant period of time, particularly when converting landsnot currently under annual crop production Set-aside lands, such asthose enrolled in the Conservation Reserve Program (CRP), maysequester more C than would their conversion to corn for ethanol for aslong as four decades, while conversion to cellulosic biofuels may nothave this effect (Piñeiro et al., 2009) As such, perennial feedstocksgrown on marginal or abandoned land may be preferred over theconversion of natural ecosystems because they may incur a lower C debt(Fargione et al., 2008).

3.2 Using land enrolled in the Conservation Reserve Program

As a response to increased demands for biofuels, increases in feedstockproduction may come from a variety of sources, including yield gainsand increases in land area under production Some of this area will includemarginal and CRP lands While models predict that much of the land usedfor perennial grass feedstocks will come from cropland, up to 31% maycome from previous CRP land (McLaughlin et al., 2002) Land underCRP is typically less productive, more erodible, and has steeper slopesthan cropland (Secchi et al., 2009) The primary objective of the CRP is

to protect lands from soil and water erosion mainly through grasslandconversion, but wildlife have also benefited from this program (Best

et al., 1997; Reynolds et al., 2001) As corn prices rise due in part tobiofuel demands, producers will likely begin to crop on marginal land,increasing water erosion, N and P losses due to sediment loss, anddecreasing soil C stocks (Secchi et al., 2009) Conversion of CRP land tobioenergy feedstock production may also create a large C debt if annualcrops are cultivated and displace the perennial species currentlyestablished on the CRP land: approximately 11 Mg C ha1 may bereleased in the first year through the conversion of CRP land to annualcrop production for biofuels, when accounting for increased GHGemissions and depletions of the soil organic carbon (SOC) stocks, whichwould take decades to repay (Gelfand et al., 2011)

Whereas corn production, particularly as a continuous corn crop, hasdeleterious impacts when grown on previous CRP lands, perennial grassfeedstocks have been suggested for growth on CRP and marginal lands

Native grasses are one of the most common choices for vegetation in CRPlands, with 2.7 Mha of CRP land under native grass plantings in 2009(CRP, 2009) These long-lived species have deep roots that can stabilizesoil and reduce water runoff, potentially furthering the goals of theCRP Several species of native grasses, in particular switchgrass, areproposed as CRP species as well as biofuel feedstock Native, warm-season grasses such as switchgrass can be suitable for both CRP andbiofuel production because they are high-yielding under low fertilizerinputs, drought tolerant, can improve soil stability, and also grow under

a wide range of conditions (Lewandowski et al., 2003) It would be ideal

if these lands could simultaneously be managed for biomass productionand also achieve the CRP objectives (Tilman et al., 2006)

The Food, Conservation, and Energy Act of 2008 permits haying andbiomass harvesting on CRP lands, provided that land is managed to protectthe soil and wildlife objectives of the program (U.S Congress, 2008) Therehave been conflicting results as to whether bioenergy cropland can providethe same services as CRP and natural grasslands, depending on thefeedstock and land use management (Fargione et al., 2009) For example,monocultures may have lower wildlife habitat quality than diversemixtures (McCoy et al., 2001) Standing biomass remaining in the fieldafter harvest may be beneficial for control of soil erosion andwater runoff, and also provide habitat for wildlife To meet both CRPand bioenergy needs, switchgrass biomass yields may be maximized withlow N fertilizer inputs and harvest after a killing frost (Mulkey

et al., 2006) Furthermore, perennial grass feedstock production onCRP land may mitigate GHG emissions by capturing a minimum of2.3 Mg CO2e ha1yr1 (Gelfand et al., 2011) However, if grasses arefertilized, increased N2O emissions may cause a net release of GHG, eventhough SOC stocks increase (Robertson et al., 2011b)

Harvest management is important for both maintaining high yields andachieving CRP goals Increased harvest frequency from one cut per year totwo or three cuts may lower yields of perennial grasses (Cuomo et al.,

1996) Even repeated annual single-cut harvests on CRP lands withoutadditional fertilizers or organic amendments may lower total yields overtime as a result of nutrient loss through plant biomass removal withoutreplacement (Mulkey et al., 2006; Venuto and Daniel, 2010) Caseswhere there have been long-term high yields on marginal lands with lowinputs, such as the diversity experiment by Tilman et al (2006), may besuspect in a bioenergy cropping system because of management choices,

as Tilman et al.'s plots were burned rather than the biomass removed.Stand persistence and productivity may impact wildlife habitat quality aswell While CRP and marginal lands have the potential to providebioenergy without interfering with food crop production, they must becarefully monitored to ensure plant vigor and persistence

3.3 Biofuels and restoration of degraded lands

The CRP enrolls marginal cropland, but lands may be degraded for otherreasons Degraded or marginal lands have low productivity and environ-mental quality due to a history of intensive use and disturbances such asfarming, mining, or erosion These lands are generally infertile and havelow SOC stocks Bioenergy feedstock such as perennial grasses and treeshave been proposed for use on marginal lands as a method to improvesoil quality, enhance SOC stocks, and improve soil fertility (Blanco-Canqui, 2010) Perennial species are already used in riparian buffers tocapture agricultural runoff Bioenergy feedstock present opportunities toaddress both ecological and energy problems “Extremophile energycrops,” species capable of high productivity under stresses such as salinity,drought, or extreme temperatures, have also been suggested for use onmarginal lands, as they are typically efficient in their use of nutrients andwater (Bressan et al., 2011)

Several biofuel feedstock options are tolerant of acid soils and varioustoxic compounds, and may aid in phytoremediation of contaminated soils(Blanco-Canqui, 2010; Peterson et al., 1998; Rockwood et al., 2004).Short-rotation woody crops (SRWC) may be able to take up nutrientsfrom wastewater, simultaneously reducing fertilizer needs and improvingwater quality (Adegbidi et al., 2001) Poplars grown as a vegetative filterstrip with landfill leachate take up 159 kg of leachate kg1 abovegroundpoplar biomass, can treat 338 kg N ha1yr1, and may yield nearly

12 Mg ha1 woody biomass (Licht and Isebrands, 2005) Based on anLCA comparing bioremediation with willows against a traditionalexcavation-and-fill method for a landfill, remediation using plants mayhave fewer negative environmental impacts (Suer and Andersson-Sköld,

2011) Giant reed (Arundo donax L.) can tolerate arsenic and heavy metalsand could be used in phytoremediation of contaminated soils (Mirza

et al., 2010; Papazoglou et al., 2005)

Highly productive perennial species tolerant of poor soil conditions mayalso aid in revegetating degraded mine soils Surface mine soils typicallycontain high amounts of rock fragments, low nutrient concentrations,and low SOC concentrations (Akala and Lal, 2000; Bendfeldt et al.,2001; Haering et al., 2004) In addition, these soils may also be compactedand acidic, all of which can make the establishment of persistent vegetativecover challenging (Haering et al., 2004) However, these degraded soilsmust be reclaimed and revegetated in order to limit soil erosion andrestore soil conditions, a process required by the Surface Mining Controland Reclamation Act (SMCRA) of 1977 Switchgrass and reedcanarygrass (Phalaris arundinacea L.) can be dominant species on surfacemine soils in Virginia, producing 16.5 Mg ha1 and 5.0 Mg ha1,respectively (Evanylo et al., 2005) Likewise, hybrid poplar has also been

recommended to reforest reclaimed mine soils as a result of high survivaland growth rates (Casselman et al., 2006) If biofuels can be grown oncontaminated or degraded land, they may be able to produce biomass forenergy without competing with food crops for land and maysimultaneously improve the condition of degraded or contaminated sites.While growing perennial feedstock on degraded lands may be beneficialfor soil and water quality, the ecological consequences of using degradedlands to produce bioenergy is uncertain Bioenergy production from aban-doned agricultural lands could provide as much as 8% of global energyneeds (Campbell et al., 2008), and including bioenergy production fromother marginal lands would further increase this value However, yieldsare generally lower on marginal lands, requiring higher selling prices forproducers to break even (Mooney et al., 2009) As with CRP land,feedstock production on other marginal lands may require greater inputs

to produce sufficient yields and may result in greater erosion andnutrient runoff (Dominguez-Faus et al., 2009) Harvesting degraded landsfor bioenergy could also have serious implications for conservation issues.While degraded lands may have lower diversity than natural systems, inmany situations they are more diverse than plantations and agriculturallands (Plieninger and Gaertner, 2011) Plant and animal species diversitymay be reduced if degraded lands are used for bioenergy crops instead ofrestored for nature conservation

4 Soil Erosion and Water Quality

Land use choices and water quality are closely linked Large-scale landuse changes from annual to perennial crops will affect sediment losses,nutrient runoff, and water yields (Schilling et al., 2008) Soil erosion andnutrient runoff into bodies of water may have serious effects on soilquality and plant productivity, as well as on ecosystem and humanhealth The impacts that bioenergy cropping systems have on soil erosionand water quality could be either positive or negative, depending on thespecies and management practices Fertilized annual cropping systems areexpected to have more negative effects on soil and water than low-inputperennial systems Increased corn production for ethanol may increasenutrient runoff due to greater use of fertilizers, particularly in large corn-growing regions such as the Mississippi River Basin An estimated 80%

of the increased production of corn will occur in the Mississippi RiverBasin, and could increase N and P loads by 37% and 25%, respectively(Simpson et al., 2008) With a scenario of 136 billion liters of corn-grainethanol by 2022, dissolved inorganic N export into the Gulf of Mexico

would increase by 34%, dramatically increasing the risk of hypoxia andcontributing to the “Dead Zone” in the Gulf of Mexico (Donner andKucharik, 2008).

Perennial species provide an opportunity to reduce the negative soil andwater impacts that annual bioenergy crops present Riparian buffer stripsestablished with grasses or trees reduce nutrient runoff, sediment loss,filterpesticides lost from crop fields, stabilize stream banks, and reduce bankerosion (Lyons et al., 2000; Udawatta et al., 2002) Perennial grasses such

as switchgrass have lower N requirements, are more efficient at using N,may reduce runoff, and also improve soil quality (Parrish and Fike,

2005) Compared to a corn-soybean rotation system, switchgrassplantings can reduce N and P runoff by 50e90%, while delaying harvestuntil winter could prevent additional nutrient runoff (Simpson et al.,

2008) Long term switchgrass stands fertilized at 68 kg N ha1 may have

N losses through surface and ground water of under 2 kg ha1, which isapproximately 3% of the N applied (Sarkar et al., 2011) Fertilizerapplication rates may affect the size of reduction of NO3eN in surfacerunoff, with greater reductions as application amount decreases: when

224 kg N ha1 is applied, there is a 16% reduction in NO3eN runoffcompared to the baseline cropping system, but with no fertilization, thisreduction increases to 65% (Nelson et al., 2006) Large-scale changes inland use to bioenergy plantations of either corn or perennial crops wouldsignificantly affect whether N or P losses would increase or decrease.Compared to a scenario where land is used for large-scale cornproduction, large-scale switchgrass bioenergy plantings would result in

a 57% decrease in NO3eN losses and nearly a 98% decrease in P losses(Schilling et al., 2008) Woody species grown for bioenergy also have thepotential to significantly reduce N and P losses from runoff (Thornton

chil-(Christian and Riche, 1998; McIsaac et al., 2010) Perennial grasses mayhave reduced nitrate losses for several reasons, including requiring lower

N applications, having a longer growing season and having a larger rootsystem, both of which should increase N uptake Diverse perennialstands, with a potential for higher root biomass and more efficientresource use, may also reduce nitrate leaching (Scherer-Lorenzen et al.,

2003)

Sediment loss from agricultural land and entry into surface waters is

a major problem Sediments can increase water turbidity, reducing plantphotosynthetic activity and productivity, and can also negatively affectfish and other aquatic invertebrates: even low levels of turbidity maydecrease primary productivity by as much as 13% (Ryan, 1991) Insugarcane production systems, residue is frequently burned on the fields

to facilitate harvest and transport, a practice that releases GHG into theatmosphere and leaves bare ground at risk of soil erosion rates of17e505 Mg ha1yr1 (Hartemink, 2008) Development of newharvesting equipment so that burning is not required can retain this soiland allow the soil to store an additional 1500 kg C ha1yr1 (Galdos

et al., 2010) Sediment loads are predicted to increase with theexpansion of annual row crops and be reduced for perennial grasses, butall feedstock plantations may generate sediment losses if grown onmarginal or steep lands (Love and Nejadhashemi, 2011) Althoughsediment and runoff from switchgrass may initially be higher during theestablishment year (Thornton et al., 1998), switchgrass may reducesediment yield by more than 99%, edge-of-field erosion by 98%, andsurface runoff by 55%, when compared to cropping systems (Nelson

et al., 2006) In contrast, SRWC may reduce sediment losses by asmuch as 85%, even in the first year of establishment (Thornton et al.,

1998)

The conversion of large areas of land to bioenergy plantations mayimpact landscape hydrology Changes in water yield would alter theamount of potential nutrient runoff, as increases in water yields may alsoincrease N and P runoff losses Shifts to large-scale production of cornfor ethanol under a large area conversion from CRP grassland to cornwould reduce evapotranspiration, increasing water flow by as much as8%, particularly during spring and late fall (Schilling et al., 2008) Incontrast, perennial species have higher amounts of evapotranspirationstemming from a longer growing period and higher biomass productionthat would reduce the water yield Under a large area conversion toswitchgrass, water yields could decrease by as much as 28%, whencompared to current cropping conditions (Schilling et al., 2008) Soilsunder CRP land have lower plant available water (PAW) within the top1.5e2 m, when compared to row crops due to greater water use by theperennial species (Randall et al., 1997)

Although conversion from annual to perennial crops may alterhydrology, each perennial feedstock species may behave differently Thedifferences in root distribution and water capture can affect changes insoil water: soil under miscanthus, with most roots in the top 0.35 m, tends

to have higher summer soil moisture at lower depths than soil underswitchgrass or giant reed (Monti and Zatta, 2009) Just as with perennialgrasses, SRWC plantations would use more water than annual crops,lowering soil moisture and potentially reducing water yield and the size

offloods if planted on a large scale (Perry et al., 2001) Compared to soilunder switchgrass, soil under miscanthus has lower soil moisture laterduring the growing season and in combination with greaterevapotranspiration, may potentially reduce surface flows by 32% ifplanted over a large area (McIsaac et al., 2010) Modeling simulationssuggest that miscanthus may be planted at under 10% of US crop areawithout impacts on hydrology; however, given that energy efficiency ismaximized by concentrating biofuel plantations, miscanthus landcoverage would likely exceed 25% or even 50% in areas surroundingrefineries or energy plants (VanLoocke et al., 2010) Reductions in wateryield may provide benefits such as reduced flooding, but may also haveadverse effects, such as later recharges of soil moisture and extendedperiods of low waterflow

4.1 Use of agricultural residues

Agricultural residues have many uses; i.e., as an input of organic matter(OM) to soils, as feed for animals, feedstock for bioenergy, and as raw mate-rials for industry Using corn stover for cellulosic ethanol may reduce SOCstocks, soil quality, crop yields, and soil faunal activities due to the loss of

OM and surface residues (Blanco-Canqui and Lal, 2007; Karlen et al.,1994; Karlen et al., 2009; Lal, 2009) The data in Figs 3 and 4 show therelationship between agricultural residues and SOC stocks or soil erosion,respectively: as residue is removed, SOC generally decreases and soilerosion increases In addition, decreases in crop residues may also reducecrop yields: for each 1 Mg ha1 of residue that is removed, grain yielddeclines by 0.10 Mg ha1 while future residue production decreases by0.30 Mg ha1 (Wilhelm et al., 1986) Agricultural residues are important

as a method for controlling nutrient runoff and soil erosion Removal of50% of crop residues can double rates of soil erosion, while higher rates

of removal can greatly increase N and P sediment losses (Blanco-Canqui

et al., 2009) Although complete removal of residues is detrimental to soiland water quality, it is possible that lower levels of residue removal,estimated at 25% or less, may be acceptable for maintaining soil andwater quality (Blanco-Canqui and Lal, 2007; Blanco-Canqui et al., 2009).The benefits of additional ethanol yield from agricultural byproducts must

be carefully weighed against the costs of increasing fertilizer inputs andpotential loss in soil quality.

1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0

Maskina et al (1993)

also includes the internal translocation of nutrients across seasons and theplant litter dynamics Internal cycling, plant material decomposition, andmineralization by microorganisms provide readily accessible sources of Nfor plants, while some N turns into stable humus compounds that provide

a slow release of N (Clark, 1977) The schematic in Fig 5 outlines thegeneral movement of N in traditional cropping systems and a perennialbiofuel system While N cycling in traditional systems such as corn andwheat is more understood, knowledge of the movement of N through

a perennial bioenergy system is incomplete and requires more work.Understanding the total system balance for N and other nutrients isimportant to optimize fertilizer applications to ensure adequate plantnutrient supply and to minimize nutrient loss Ideally, biofuel specieswould be productive under low nutrient inputs and would furtherreduce their nutrient needs through three factors: nutrient-poor harvestedtissue, internal movement of nutrients to roots, and litter decomposition(Tom, 1994)

5.1 Nitrogen and litter/residue management

Leaves, agricultural residues, and other sources of plant litter are an tant part of nutrient cycling In a mixed-grass prairie, litter may comprise up

impor-to 60% of the aboveground biomass and contain approximately 50% and60% of aboveground C and N stocks (Schuman et al., 1999) Litterdecomposition rates can be affected by its quality, as well as litter N andlignin concentrations (Fog, 1988; Melillo et al., 1982) Litter rich in Nand low in lignin content tend to decompose more quickly and maymake N available for plant uptake more rapidly (Melillo et al., 1982).Leaf litter impacts nutrient cycling through incorporation into the soil,through microorganism activities, and by leaching of surface litter (Clark,

1977) As willow stands establish, N release from leaf litter may reachover 100 kg N ha1, while internal recycling may be over 50 kg N ha1,which will significantly reduce N fertilizer needs (Tom, 1994).Nitrogen cycling within an ecosystem may also be measured by using

15N-enriched fertilizers and tracking it within plant parts and soil In a

Figure 5 Movement of nitrogen within a) traditional cropping and b) perennial fuel systems Percentages represent the amount of N added to the system that is removed by each component Note: for part b, percentages do not add up to 100 due to values taken from multiple sources, uncertainty in calculations, and uncalculated N translocation and storage in roots Data from Adler et al (2007); Bransby et al (1998); Liebig et al (2006); Marshall et al (1999); McLaughlin et al (2002) , and Smil (1999)

bio-shortgrass prairie, 65% of applied15N was in plant material in thefirst yearand 55% remained in plant tissue by the fifth year due to the cyclingprocesses inherent in the ecosystem (Clark, 1977).

Crop residues may also impact soil N as well as N uptake by plants as N

in residues mineralizes and becomes available for plant use (Maskina et al.,

1993) Applying 16 Mg corn residue ha1yr1 can increase soil organic N(SON) by as much as 37% over an 11-year period when compared to anapplication rate of 2 Mg ha1yr1 (Larson et al., 1972) As residuesdecompose N is released, which may be immobilized by microorganisms,taken up by plants, or lost through nitrate leaching and N2O formation(Aulakh et al., 1991; Maskina et al., 1993) If residues are removed,additional increases in fertilizers may be required to compensate for thenutrients removed in stover biomass (Karlen et al., 2011) Therefore, theloss of leaf litter and surface residues for use as biofuel feedstock may havesignificant impacts on nutrient return to the system and nutrient cycling

5.2 Nitrogen uptake and biomass removal

In addition to losses through runoff, leaching, and volatilization, nutrientsare also removed from the system through biomass harvest As such, peren-nial plants and warm-season species that are efficient in their nutrient use orcan translocate nutrients to belowground storage organs will require lessfertilizer The data in Table 3 show estimates of some of the losses in

Table 3 Nitrogen losses in biofuel cropping systems

Crop

N content in harvested biomass (kg Mg L1 )

Total N harvesteda(kg ha L1 )

Nitrate leaching (kg ha L1 )

N runoff (kg ha L1 )

a Based on biomass yields from Table 1 and biomass N content (kg Mg1) ( Thornton et al., 1998 ).

b Leaching and runoff for no-till corn system.

c

Measured leaching under a high N application rate of 140 kg N ha1.

d

Assumes stover removal of 25%.

Sources: Beale and Long (1997); Bransby et al (1998); Christian et al (2006); Lemus et al (2008); Maskina et al (1993); McIsaac et al (2010); McLaughlin et al (2002); Mortensen et al (1998); Partala

et al (2001); Randall et al (1997); Thornton et al (1998); Tom (1994); Wrobel et al (2009).

different biofuel systems, and indicate that the N content in perennial bioenergycrops is only a fraction of what is in corn grain Nitrogen concentrations inharvested willow stems may be as low as 4 kg N Mg1 (Tom, 1994).

In miscanthus, N in aboveground tissues at harvest is similarly low, at

5 kg N Mg1 (Beale and Long, 1997) Nitrogen concentrations in reedcanarygrass and switchgrass may be slightly higher, at 6e8 kg N Mg1 and8e10 kg N Mg1, respectively (Bransby et al., 1998; Christian et al., 2006;Lemus et al., 2008; Partala et al., 2001) In contrast, corn grain contains13e14 kg N Mg1 and stover 4e6 kg N Mg1(Maskina et al., 1993) Theability for plants to take up and use fertilizers also varies among species:fertilized switchgrass may have an N recovery of nearly 66%, while cornrecovery is estimated at approximately 50% (Bransby et al., 1998)

Although perennial species generally have lower concentrations of N inharvested biomass, some N still remains in the tissue Some of this remain-ing N may also be recovered through nutrient recovery during the bioen-ergy process As much as 78% and 23% of total N fertilizer input may berecovered as NH3 from switchgrass and corn stover bioenergy systems,respectively, which could potentially be applied back to the croppingsystem (Anex et al., 2007)

5.3 Gaseous emissions and volatilization

Gaseous soil N emissions are another source of system N loss and can be

a significant source of GHGs Emission of nitrous oxide (N2O) can bedue to direct factors such as the rate and type of fertilizer use, organicamendments, crop species, and crop residue management, or indirect sour-ces such as leaching and runoff, and atmospheric deposition of Ncompounds (Mosier et al., 1998) Estimates of N2O emissions caused by

N applications vary and may be as high as 8% of N applied, althoughmost estimate emissions at 5% or less (Crutzen et al., 2008; IPCC, 2006;Mosier et al., 1996) Direct N2O emissions are lower for switchgrass andpoplar bioenergy cropping systems and higher for reed canarygrassfeedstock and conventional cropping rotations (Adler et al., 2007).Perennial species generally have lower N2O emissions than annualspecies as a result of lower N inputs, greater N use efficiency, higher soilwater content through shading, and reduced cultivation activities (Kavdir

et al., 2008) In addition, emissions from unfertilized switchgrass areexpected to be less than those from CRP grassland, as plant material onthe unharvested CRP lands decomposes and is denitrified (Chamberlain

et al., 2011)

In many cases, complete nutrient balances are lacking for biofuel ping systems Theoretically, perennial feedstock should have more efficientnutrient cycling, require fewer inputs, and retain nutrients within thesystem Even though perennials may require fewer inputs, harvesting

crop-biomass removes some nutrients from the system which must be replaced ifyields are to be maintained over the long term However, it is unclearexactly how fertilization affects yields of many feedstock, nor is it knownhow much remains in plant tissue, in the soil, or is lost through runoff,leaching or volatilization As such, more information is needed on thenutrient balances of biofuel crops.

6 Human Impacts on Biodiversity

Humans have impacted ecosystems over millennia The entire planethas been affected by human activities that have resulted in land transforma-tions, changes in biota and genetics, and alteration of biogeochemical cycles(Haberl et al., 2007; Leu et al., 2008; Vitousek et al., 1997) In the pastseveral decades, the impact that biodiversity may have on ecosystem func-tioning has been a subject of intense debate for scientists (Cardinale et al.,2006; Huston et al., 2000; Loreau et al., 2001), as well as how ecosystemfunctioning may change due to human-caused species extinctions andintroductions (Balvanera et al., 2006; Hooper et al., 2005; Naeem et al.,2009) Plant biodiversity has been associated with many other ecosystemproperties such as system stability and invasibility (Tilman, 1999), butdiversity at multiple trophic levels may also impact ecosystem functioning(Naeem et al., 1994) Plant biodiversity in agricultural settings can affect

a plethora of ecosystem services, including enhancing nutrient cycling,reducing pest occurrences, enhancing habitat for pollinators, andreducing weed competition (Altieri et al., 1983) Some biofuel systemspresent an opportunity to improve diversity in agricultural settings,reduce the land area allocated to traditional monoculture crops, andpotentially alter ecosystem properties and services

6.1 Biodiversity in agroecosystems

The initial research on biofuel crops focused on fertilized monocultures(e.g., corn, sugarcane, soybeans, switchgrass plantations) Monoculturesare typically easier to establish and maintain, and harvest may be simpler

as well, since all plants within thefield mature at approximately the sametime However, the high cost for fertilizers, pesticides, and water inputs

to achieve high yields may be a disadvantage in monoculture croppingsystems (Williams et al., 2009) As resources become more limited,producers are increasingly attempting to simultaneously achieve multipleobjectives such as energy production, C sequestration, wildlife habitatenhancement, soil and water protection, and forage production

Incorporating diversity into bioenergy systems is a potential answer tothis problem Plant diversity is often associated with higher biomass yieldsand greater nutrient use efficiency (van Ruijven and Berendse, 2005),but in several cases no correlation between diversity and NPP has beenobserved (Huston et al., 2000) Diversity is already implemented inpasture settings, where incorporating legumes and grasses improves foragequality (Bullock et al., 2001; Sleugh et al., 2000), and species-richpastures can produce yields up to 60% greater than species-poor pastures(Bullock et al., 2001) The impact that plant diversity may have onbioenergy production and yields is less clear, however, and in low-inputsystems, careful selection of a few well-adapted species can produceyields similar to those of diverse mixtures (DeHaan et al., 2010).

6.2 Diverse perennial grasslands

Low-input high-diversity (LIHD) perennial grasslands have been considered

by some as a system that can potentially be high-yielding while requiringlower amounts of water, fertilizer, and pesticides (Tilman et al., 2006).Weed management may also be reduced in diverse grasslands, which maysuppress weedy species (Picasso et al., 2008) Diverse grasslands haveadditional ecological benefits, such as increased stability, resilience, andmore rapid recoveries from environmental perturbation (Hector et al.,2010; Tilman and Downing, 1994) Nutrient cycling is also affected byplant diversity and community composition In plots with legumes, nitrateleaching is negatively associated with diversity and legume abundance(Scherer-Lorenzen et al., 2003) Diverse grasslands also store up to 600%more soil N and 500% more soil C than monocultures due in part toincreased root biomass (Fornara and Tilman, 2008) However, the increases

in soil C cannot be attributed entirely to root biomass, but also to plantdiversity (Steinbeiss et al., 2008) Total (root and soil) C sequestration inLIHD plantings may reach 4.4 Mg C ha1yr1 in the first decade anddecrease to 3.3 Mg C ha1yr1after that, compared to monocultures thatmay store only 0.1 Mg C ha1yr1in thefirst decade and essentially none

in later years (Tilman et al., 2006), although in some cases moderatelyfertilized switchgrass monocultures can store more C than native prairiemixtures (Zeri et al., 2011)

Biomass production is just one step in the bioenergy system pathway.Plant material must then be converted into energy as efficiently as possible,and the effects of diversity on ethanol yields are debatable Perennial diver-sity is often associated with increased biomass and ethanol yields (Tilman

et al., 2006), but negative relationships have also been observed (Adler

et al., 2009) Energy production from biomass may vary due to thechemical compositions of species within mixtures as well as potentialallelopathy due to neighboring species (Adler et al., 2009; Carroll and

Somerville, 2009) Where there is a negative diversity-ethanol relationship,increasing diversity reduces ethanol yields due to two factors: lower plantbiomass yields, and lower cellulose and hemicellulose contents associatedwith higher species richness, such that the greatest ethanol yields comesfrom grasslands dominated by warm-season prairie grasses (Adler et al.,

2009) Diverse plantings with few inputs may not maximize ethanolyields, as switchgrass monocultures with moderate N fertilization(74 kg N ha1) can produce 93% more energy than LIHD grasslands(Schmer et al., 2008) The relationship between diversity and energyyields is a field of study that requires further research

6.3 Effects on wildlife

Feedstock diversity at both small and large scales is important for wildlifeuse and biocontrol Conversion of natural ecosystems to biofuel plantationsmay cause declines in floral and faunal diversity, with large losses ofspecialist species and species of concern (Danielsen et al., 2009; Fletcher

et al., 2011) Large-scale plantations may actually generate biotic ties that are distinct from other natural and agricultural habitats; forexample, natural and SRWC plantations do not support the same birdcommunity (Christian et al., 1998) Compared to perennial grasslands,corn biofuel fields typically contain lower levels of insect species richnessand abundances, as well as avian diversity, yet management of perennialgrasses for biofuels may lower wildlife habitat value (Gardiner et al.,2010; Robertson et al., 2011a) The conversion of cropland toswitchgrass plantations may increase habitat of some grassland bird species

communi-of concern (Murray et al., 2003) Bird diversity is generally higher inwarm-season grassland hayfields than in cool-season hayfields becausewarm-season fields are left undisturbed during the nesting season and

a 15e20 cm stubble remains on the field during the winter to providecover (Giuliano and Daves, 2002) Perennial bioenergy grasslands couldprovide similar benefits to wildlife, since biomass is typically harvestedonce or twice a year and is cut to a 15e20 cm height Increased plantdiversity may increase wildlife use, since diverse perennial bioenergysystems create a more heterogeneous habitat and have less soildisturbance than annual monocultures (Gardiner et al., 2010)

Heterogeneity is relatively low on commercial plantations, which mayreduce habitat quality (Christian et al., 1994) Managing switchgrass areas tomaintain vegetation structural heterogeneity by leaving some fieldsunharvested would further improve wildlife habitat quality (Murray et al.,

2003) At the landscape level, bioenergy plantations could affect wildlifemovement and population processes if feedstock plantations serve aspopulation sinks that attract animals but are not areas that produce highreproduction rates (Christian et al., 1994) When managing diverse

grasslands for biofuel, trade-offs between wildlife habitat quality andpotential decreases in ethanol yield may also have to be considered(Adler et al., 2009).

Although wildlife use of native, warm-season grasses and corn arewidely studied, the effects of other perennial grasses and trees have alsobeen researched Based on animal diversity, SRWC plantations can provide

as good a habitat as croplands, but are poorer than natural woodlands, withspecies present commonly being habitat generalists rather than obligateforest species (Christian et al., 1998) Both miscanthus and reedcanarygrass provide habitat and cover from predators for small mammals(Semere and Slater, 2007) However, the tall and dense growth of reedcanarygrass monocultures create a poor breeding habitat for skylarks inEurope (Vepsäläinen, 2010) More bird species were found within young(2e3 year old) and more open miscanthus plantings than dense reedcanarygrass fields, although reed canarygrass field edges may support

a larger number of bird species (Semere and Slater, 2007) Althoughwildlife use of miscanthus in the US is not well documented, its value tograssland birds is suspect due to plant architecture that is different frommost plants in the US (Fargione, 2010) As the acreage dedicated tolignocellulosic feedstock increases, more research is needed to determinethe wildlife impacts of large-scale plantings of herbaceous and woodybioenergy crops

6.4 Diversity at the landscape level

Wildlife use may be affected by feedstocks at both the field level, asdescribed above, and at the landscape level A high concentration of nativeperennial grasslands for bioenergy around refineries may be advantageous tonesting and obligate grassland bird species (Robertson et al., 2011a).Conversely, this clustering of biofuel plantations in one area will alsoreduce landscape heterogeneity and may reduce biotic diversity (Wiens

et al., 2011) Establishment of SRWC on a large scale is predicted todecrease the overall biodiversity, with the largest declines in reptiles,butterflies, and birds, although vascular plant diversity is predicted toslightly benefit (Louette et al., 2010) Landscape level changes in diversitymay also impact pollination activities, as large, dense, plantations ofoilseed rape (Brassica napus L.) may increase competition for pollinatorsbetween crop plants and natural plants, reducing pollinator visits and seedset within nearby natural grasslands (Holzschuh et al., 2011)

Habitat heterogeneity at several scales is an important factor in taining species diversity (Hutchings et al., 2000) Wildlife may benefitfrom leaving some fields unharvested, which would increase landscapeheterogeneity and provide habitat for more species (Roth et al., 2005) Intree plantations, maintaining stands of varying ages through a region

main-would increase structural variability and improve wildlife diversity(Hanowski et al., 1997) Maintaining a mosaic landscape that containsboth open and forested areas may enhance avian diversity in willowplantations (Berg, 2002) The diversity and wildlife potential ofbioenergy systems may not be maximized if they are managed intensively

as monocultures However, if managed properly, with heterogeneity atmultiple levels (i.e., field, farm, and landscape) in mind, bioenergy-cropping systems may enhance wildlife diversity

6.5 Pests and biocontrol

Changes in plant diversity have additional impacts on biocontrol, wherenatural enemies keep pest populations in check Biocontrol is an importantecosystem service that is valued at $417 billion per year globally and has

a value of $24 ha1 of cropland (Costanza et al., 1997) However, theintensification and simplification of agricultural landscapes has increasedthe odds of a severe spread of pests or disease The loss of speciesdiversity and genetic diversity in agricultural settings has increased plantsusceptibility and the rate of pathogen and pest spread (Gonzalez-Hernandez et al., 2009) Crop monocultures have a higher risk of severepathogen attacks and pest infestations that could reduce yields or elserequire substantial pesticide inputs (Landis and Werling, 2010; Mitchell

et al., 2002) Increasing plant diversity reduces the chance that

a pathogen will encounter a susceptible host and reduces the chance of

a pest outbreak decimating an entire system

Bioenergy systems are no exception If grown in monocultures, a severedisease or insect attack can significantly reduce productivity and yields(Hartman et al., 2011) In contrast, diverse bioenergy plantations of nativeperennial species, already suggested as a potential method of increasingproductivity, may also be beneficial for reducing the severity andfrequency of pests Compared to corn, biocontrol may be greater ingrasslands with moderate diversity and floral species richness, whichincrease natural enemy populations and the predation of pest eggs(Werling et al., 2011) Similarly, a genetically diverse plantation may alsoprovide some biocontrol by requiring pests to search farther for susceptiblehosts, slowing their spread across an area (Peacock et al., 1999) Therefore,managing for species and genetic diversity may become an important way

to keep pesticide inputs low, minimize pesticide impacts to theenvironment, and maintain profits

Interactions between agronomic pests and feedstock species must also beassessed at the landscape level Biofuel feedstock can be a host for both bene-ficial insects and agricultural pests, as well being reservoirs for insect-vectoredplant viruses (Landis and Werling, 2010) Landscape composition is an

important aspect of biocontrol services, with increasing biocontrol asperennial habitat increases (Werling et al., 2011) As more land is planted

to corn, biocontrol by predators of crop pests may decrease due to the loss

of landscape diversity (Landis et al., 2008) The conversion of marginal andnon-crop lands to biofuel production may also reduce the population ofnatural enemies to pests, increasing pest populations (Tscharntke et al.,

2005) Conversely, agricultural landscapes with higher diversity (i.e.,cropland interspersed with grassland and forestland) may increase theabundance of some native beetles that prey on crop pests (Gardiner et al.,

2009) Biocontrol may therefore be maximized by including perennialgrass and woody species plantations in an agricultural landscape

While perennial species such as switchgrass have few insect pests, it ispossible that pests have yet to be identified or that management of biofuelmonocultures may increase the prevalence of pest damage (Parrish andFike, 2005; Prasifka et al., 2009) Recently, several fungal diseases havebeen identified on switchgrass, suggesting that it may be a host for otherdiseases and that pathogens may need to be managed to sustain biomass(Layton and Bergstrom, 2011; Vu et al., 2011) Fall armyworm (Spodopterafrugiperda ( J.E Smith) (Lepidoptera: Noctuidae)), a pest to corn, can survive

on switchgrass and miscanthus under ideal greenhouse conditions, butsurvival is poor to none in the field, suggesting that biofuel suitability forthis pest may be low (Nabity et al., 2011) Similarly, two species ofaphids have been reported on miscanthus stands in four US states,suggesting that damage from aphids and transmitted diseases is possibleand warrants more evaluation (Bradshaw et al., 2010) Cultivar selectionmay also increase pest susceptibility if selections that increase agronomicvalue (productivity, nutritive value) come at the expense of hostresistance In switchgrass, highly selected cultivars exhibit greatersusceptibility to aphid-transmitted viruses and a greater preference byaphids, which may be linked to increased productivity (Schrotenboer

et al., 2011)

The introduction of exotic species may have unexpected consequences

to current ecological and agronomic situations For example, it is possiblethat miscanthus could either slow the buildup of a Bt resistant westerncorn rootworm (Diabrotica virgifera virgifera LeConte; WCR) population

or alternatively could be a reservoir for WCR population increases thatmay threaten corn production (Spencer and Raghu, 2009) Exoticfeedstocks are not the only potential pest source: native switchgrass standsmay also be reservoirs for insect-transmitted viruses, such as barley yellowdwarf viruses (BYDVs), that infect cereal grain crops and can reduceyields by 15e25% (Lister and Ranieri, 1995; Schrotenboer et al., 2011).The interactions between pests and large-scale bioenergy plantations isnot well understood and need be more thoroughly examined

7 Biofuels and the Soil Carbon Budget

One of the primary goals of biofuel feedstock is to serve as C-neutral

or even C-negative sources of energy, mainly through C sequestration TheSOC stocks either increase or decrease, depending on land use changes,feedstock choices, and management practices (Conant et al., 2001; Guoand Gifford, 2002; Tolbert et al., 2002) The SOC is an important compo-nent of soil and environmental quality It influences both soil physical prop-erties (e.g., soil color, structure, stability, and water-holding capacity) aswell as soil chemical properties (e.g., cation exchange capacity, pH buff-ering, and mineral decomposition) (Brady and Weil, 2008) Thus,changes in SOC stock may impact agriculture and the environment byaltering plant productivity and soil and water quality

7.1 Land/soil preparation

Land for biofuel crops may be allocated from several sources, includingpreviously cropped land or by conversion from natural ecosystems As dis-cussed inSection 3.1, LUC can result in significant losses in SOC stock and

a large C debt Land conversion may result in losses of 50e70% of theinitial C stock through mineralization, soil erosion, and dissolved organiccarbon (DOC) leaching (Lal, 2001) The SOC sequestration levels areestimated to be much greater in perennial species but may also be site-specific (Blanco-Canqui et al., 2005) Even if natural areas are neitherburned nor tilled in preparation for biofuel establishment, soil propertiesmay still be affected: conversion from native perennial grasslands tonever-tilled annual crops and using best management practices (BMPs)suggest that labile C fractions are reduced and soil biota are affectedwithin three years of land conversion (DuPont et al., 2010)

Besides land clearing, site preparation for biofuel crop establishmentmay include tillage, site preparation, fertilization, and herbicide applica-tions, all of which are sources of GHG emissions (Adler et al., 2007) andare described in Table 4 Conventional tillage (CT), which thoroughlyturns over the soil to leave less than 15% residue cover, can lead tooxidation of OM and exacerbate erosion by wind and water In contrast,conservation tillage practices such as reduced till or no-till (NT) leavemore agricultural residue on the surface and may reduce SOC losses byerosion and mineralization (Lal, 1998) Conversion from CT to NT maysequester up to an additional 0.57 Mg C ha1 until the system reaches

a new equilibrium in about 20 years (West and Post, 2002) In addition

to SOC sequestration, GHG emissions from farm machinery are reducedwhen comparing CT and NT systems: NT corn reduces CO2 emissions

by about 20%, while NT for switchgrass establishment can reduce CO2emissions by nearly 50% (Adler et al., 2007) A consequence of reducedmachinery use in NT systems is increased herbicide and pesticide inputs,but in combination with enhanced SOC sequestration and loweredmachinery costs (Lal, 2004c), NT systems have the potential for short-term sequestration of as much as 200 kg C ha1 in annual crops, whilereducing agricultural inputs by 31 kg C ha1indefinitely through changes

in agricultural practices (West and Marland, 2002)

7.2 Soil carbon budget

The soil C stock, containing approximately 2500 Pg C globally to 1-mdepth (Lal, 2004b), is closely connected to the biotic and atmospheric Cstocks primarily through photosynthesis and respiration (Fig 6), but alsothrough the anthropogenic impacts of deforestation and agricultural andurban activities (Kirschbaum, 2000; Lal, 2004b) There are twocomponents to the soil C stock: organic and inorganic The SOC stock

is estimated at 1500 Pg within the top 1-m soil and 2344 Pg, to a 3-mdepth (Jobbágy and Jackson, 2000) Almost 4.5 times larger than thebiotic C pool, the soil C stock presents an opportunity to sequesteratmospheric CO2 through biofuel feedstock production (Lal, 2004b).However, determining the net balance of C during biofuel productionrequires an accurate accounting of changes in the soil C stock Althoughthere is a large potential of sequestering a significant amount of C underbiofuel feedstock, this potential is also associated with a large amount ofuncertainty that may make precise calculations of C balances highlychallenging (Ney and Schnoor, 2002)

Table 4 Emissions from select farming operations, in terms of C equivalents (CE).

Feedstock production and management choices affect the SOC budget(Conant et al., 2001; Lemus and Lal, 2005) Compared to CT corn, theSOC stock under NT corn may be nearly 25% larger, while on averageSOC stocks under corn are only two-thirds of that under woodlands, sug-gesting that there is potential to store more SOC under NT (Mishra et al.,

2010) While converting CT to NT corn may improve SOC stocks, cornhas a limited SOC sequestration capacity because it is an annual with a lowroot:shoot ratio Perennial species such as switchgrass and willow, withdeep and extensive root systems, may have root:shoot ratios of up to0.5e0.7, and have the potential to store greater amount of SOC (Zan

et al., 2001) Compared to cultivated croplands, soils under switchgrasscan store an additional 15.3 Mg SOC ha1 when measured to 120 cmdepth (Liebig et al., 2005) Choices that increase root biomass, rootingdepth, and total plant biomass tend to increase the SOC stock and loweratmospheric GHG emissions

8 Invasive Potential of Bioenergy Crop Species

There is a set desirable characteristics for a potential bioenergy crop:

a perennial nature, high NPP, high nutrient and water use efficiency, location of nutrients to belowground in fall, competitiveness against weeds,few pests and diseases, and sterility (Raghu et al., 2006) However, many of

Figure 6 Carbon flows through the atmosphere, plants, and soil that impact C sequestration in soils, modified from Lal et al.(2011 ).

these traits are also advantageous to invasive species, and as a result, severalpotential bioenergy crops are already naturalized in portions of the US orhave a closely related species that is a known invasive species (Barneyand Ditomaso, 2008).

Human activities may greatly impact the invasive potential of any duced bioenergy species Anthropogenic activities such as agriculture, trans-port, and environmental modifications have rapidly increased the rate ofnon-native species introductions, either by accident or purposefully(Pimentel et al., 2000) Of the 5800 species of plants intentionallyintroduced to the US, 128 species have become pests (Pimentel et al.,

intro-1989) One example is kudzu (Pueraria montana (Lour.) Merr var lobata(Willd.) Maesen & S Almeida), originally introduced for soil erosioncontrol, is now established on 3 Mha in the southern US and isincreasing at a rate of 50,000 ha yr1, and has reached southern Ohio(Forseth and Innis, 2004) Johnsongrass (Sorghum halepense (L.) Pers.),originally planted as a forage species, is now one of the top ten worstweeds in the world (Holm, 1969) The weedy potential of exoticbioenergy species must be analyzed in order to limit the risk ofintroducing invasive or noxious species

Protocols designed to determine the invasive potential of a species, such

as the Weed Risk Assessment (WRA)first developed in Australia, can cate whether or not a species possess a set of life history, dispersal, andhabitat characteristics along with a previous record of impacts in otherregions that may make it invasive, but even these are not completely accu-rate in their predictions (Cousens, 2008) The WRA has been used toascertain the invasive potential of three biofuel crops, switchgrass,miscanthus, and giant reed, with both switchgrass and giant reedidentified as potentially highly invasive in parts of the US (Barney andDitomaso, 2008) Tools such as the WRA and others should be usedwhen assessing the potential ecological consequences of any introducedspecies

indi-In accordance with the results of the WRA, both giant reed and reedcanarygrass are currently listed as invasive species in the US but are stillbeing considered as bioenergy feedstocks (USDA-NRCS, 2011) Napiergrass (Pennisetum purpureum Schumach.), a warm-season grass fromAfrica, is a highly productive species that can yield up to 88 Mg ha1and also has biofuel potential (Somerville et al., 2010) However, theFlorida Exotic Pest Plant Council lists Napier grass as an exotic invaderthat has naturalized and has populations expanding in native Floridahabitats (Florida Exotic Pest Plant Council, 2011) Climate change mayfurther increase the invasive potential of a species as temperatures warm,precipitation patterns shift, and the concentration of atmospheric CO2increases As temperatures increase, the suitable environment forswitchgrass is expected to expand northward, although the western US

still remains largely unfavorable due to water limitations (Barney andDiTomaso, 2010) Other species, which are currently not invasive, mayincrease in aggressiveness under the future climate conditions Thus, it isessential to examine exotic feedstock responses to potential climatechanges to prevent a possible invader from establishing.

In addition to climate change, breeding efforts could potentially impactthe invasibility of biofuel species Cultivars of reed canarygrass, used as a foragespecies, have been bred for increases in productivity and forage use, but theeffects that this may have on its invasive potential are unclear (Jakubowski

et al., 2011) However, cultivars and invader populations produce similaryields in wetland environments, suggesting that breeding is not the causefor reed canarygrass invasion (Jakubowski et al., 2011) Conversely, ifbreeding efforts can result in species being less fecund while still remainingproductive, such as the sterile triploid Miscanthus x giganteus, the invasivepotential could drop significantly (Jakob et al., 2009)

8.1 Invasive species as feedstock

As discussed in Section 8, a few potential biofuel species are alreadyinvading native US habitats Reed canarygrass, an invader frequentlyfound in wetlands of the northern US and southern Canada, canwithstand a wide range of soil moisture regimes and is one of the mosthigh-yielding cool-season grasses, especially under drought conditions(Galatowitsch et al., 1999) Similarly, giant reed, also an invader inwetlands, has a broad tolerance for a range of environmental conditions,and is highly productive (Quinn and Holt, 2008) When these speciesinvade an ecosystem, they typically become the dominant species, and as

a result, diversity-related ecosystem services may suffer Eradicatinginvasive species is a costly endeavor with no guarantee of success.However, it may be possible to harvest the biomass in areas where thesespecies have already invaded as a method of ecosystem restoration,simultaneously providing low-cost feedstock and mitigatingeutrophication through nutrient removal (Jakubowski et al., 2010) Inthis manner, an invaded ecosystem may slowly become less dominated

by an invasive biofuel species while providing income from harvestingand conversion to biofuels

9 Food versus Fuel

One of the increasingly important challenges for humans is to mine how to allocate limited resources such as land, water, and nutrients.Biofuel feedstocks, although providing energy and reducing GHG

deter-emissions, do so at the cost of diverting these limited resources from foodproduction (Pimentel et al., 2009) As a result, food crops and biofuel cropsmay compete for the use of arable lands and resources, so biofuelproduction could potentially jeopardize food security which is already athigh risk As the demand for food crops increase due to both populationgrowth and biofuel production, food prices are also expected to increase.The cost of food has been on the rise since the early 2000s, with a sharpincrease in 2008 and a lesser increase in 2010 (FAO, 2011) Between

2007 and 2008, the global price index for food increased by 45%, withbiofuel production accounting for anywhere between 3 and 30% of theprice increase, although there is disagreement about the impact thatbiofuels played in recent price increases (Ajanovic, 2011; Mueller et al.,2011) While rising food prices affect all consumers, the poor, who spendgreater than 70% of their household budget on food, are more at risk(FAO, 2011) As a result of rising food prices, the poor may faceincreasing malnutrition and reductions in caloric intake

In 2008, about 13% of the global population was undernourished, butthis number increased to 33% in the least-developed countries (FAO,

2011) Currently, global cropland area per capita is just under 0.22 ha,while a minimum of 0.5 ha per capita is necessary for a diverse andnutritious diet (Pimentel et al., 2008) In addition, the global population

is expected to rise to more than 10 billion by 2100 (Fig 2b), generatingmore pressure to increase food supplies or risk increasing food insecurity.Continued biofuel expansion into croplands and competition for primearable land with food crops may continue to reduce per capita croplandand may also result in the cultivation of marginal or abandonedagricultural lands Yields on poorer soils will be lower, or else requiregreater resource inputs, both of which could drive the price of food upfurther

Grain products and other crops used both for food and biofuel will not

be the only types of food impacted through large-scale biofuel tion Corn grain in the US is used for human consumption, as a feed grainfor livestock, for industrial purposes, and is also exported to other coun-tries In 2008, one quarter of corn grain was used for biofuel production,but with corn-based ethanol production predicted to consume as much as35e40% of the US corn crop by the end of this decade, reductions toother corn grain uses will likely occur (Mueller et al., 2011; USDA,2010) The livestock industry, which relies heavily on corn as a feed grain,may reduce meat production or increase prices, as will other food indus-tries that depend on corn (e.g., dairies and bakeries) For every $1 per

produc-25 kg ($1 bu1) increase in corn prices, the average food price is predicted

to increase by 0.8%, with a 2.9% increase in the cost of meat and a 1.7%increase in dairy products (Hayes et al., 2009) Increases in corn prices due

to biofuel demands in the US will also affect global trade and the price of

corn worldwide Biofuel production using feedstock that are also staplefoods is not a long-term solution in the search for renewable energyfrom biomass.

Lignocellulosic feedstocks may reduce some of the competition betweenbiofuels and food, depending on management choices.Tilman et al (2009)recommend establishing perennial species on abandoned cropland, carefuluse of agricultural and wood residues, double cropping of biofuel speciesand food crops, and the use of municipal wastes However, perennialgrasses and trees growing on marginal croplands, grasslands, or grazinglandsmay displace grazing livestock, further altering the supply of livestock andraising the cost of meat Grasslands would have to be managed carefully tomaintain other ecosystem services such as livestock production and wildlifehabitat, which would reduce the total amount of area for perennialfeedstock plantations Perennial feedstocks on marginal lands could supply

a portion, but not all, of energy needs

Residues are already a vital part of nutrient cycling and act to protect soilquality Removal of over 25% of residues can negatively impact crop yieldsand cause soil erosion, as mentioned inSection 4.1 Double cropping may be

a challenge with current leading feedstock species, as most are perennial andcould not work in a double cropping system, while others are food /feedstock crops (i.e., corn) that could once again bring into question thefood versus fuel debate The use of municipal wastes does hold potential

Of the more than 220 million Mg of municipal waste produced in the US

in 2009, almost 12% was burned with energy recovery (EPA, 2010) Itmay be possible to collect municipal organic waste and process it intobiofuels, which would simultaneously reduce deposits in landfills and notrequire arable land for energy production Some agricultural and treebyproducts (i.e., rice husks, coconut shells, food packing biomass, sawdust)may have specific niches as biofuel feedstocks and not compete for thefinite resources Approximately 80 million Mg of rice husks are producedannually worldwide, most of which are deposited in fields or landfills orused as fuel for cooking, but if converted to energy, could provide1.2 109

GJ each year (Natarajan et al., 1998; Tsai et al., 2004) Similarly,many other organic byproducts may also be able to be converted into energy

10 Conclusions

Bioenergy systems are increasingly called to be multi-functional andprovide energy along with other ecosystem services Life-cycle assessmentsfocusing on energy efficiency and GHG emissions are an important part ofevaluating bioenergy systems, but assessments must not stop at these twofactors In order for any bioenergy feedstock to be a viable long-term