Institute of Transportation Studies
UC Berkeley Transportation
Sustainability Research Center
( University of California, Berkeley)
Year 2007 Paper UCB - ITS - TSRC - RR - 2007 - 4
Biofuel Boundaries: Estimating the
Medium- Term Supply Potential of
Domestic Biofuels
Andrew Jones  Michael O’Hare†
Alexander Farrell‡
 UC Berkeley Transportation Sustainability Research Center
† UC Berkeley Goldman School of Public Policy
‡ UC Berkeley Transportation Sustainability Research Center
This paper is posted at the eScholarship Repository, University of California.
http:// repositories. cdlib. org/ its/ tsrc/ UCB- ITS- TSRC- RR- 2007- 4
Copyright  c 2007 by the authors.
Biofuel Boundaries: Estimating the
Medium- Term Supply Potential of
Domestic Biofuels
Abstract
We estimate the physical supply potential of biofuels from domestic mu-nicipal
solid waste, forestry residues, crops residues and energy crops grown
on existing cropland using optimistic assumptions about near- term conversion
technologies. It is technically feasible to produce a significant amount of liquid
biofuel ( equivalent to 30- 100% of 2003 gasoline demand) without reducing do-mestically
produced food and fiber crops or reducing the total calories available
as domestic animal feed. Most of this supply can be attributed to the poten-tial
of energy crops, with the combination of municipal solid waste and forestry
residues supplying between 10% and 30% of recent gasoline demand.
Our modeling approach to energy crops is unique in that it explicitly models
interactions between the feed and fuel system using an optimization procedure
that adjusts cropland allocation among major crops subject to a simple food
security constraint. Our modeling indicates that sizable increases in biofuel pro-duction
need not result in decreased availability of food or animal feed, but will
require changes in the composition of livestock diets away from hay and soymeal
toward either whole corn or feed coproducts of biofuel processing such as dis-tillers
grains. Whole corn yields very high levels of digestible calories per land
area, so shifting away from soymeal and hay to corn feed permits the same total
level of digestible calories to be produced from a smaller area. Furthermore, the
coproduction of animal feeds with biofuels relaxes the need to grow dedicated
feed crops at all. Thus, under our food security constraint, energy crops which
yield feed coproducts ( such as corn ethanol) can be grown on a larger area than
other energy crops, potentially yielding higher total levels of biofuel than other
crops ( such as switchgrass) that yield more biofuel but less animal feed per land
area. When the food security constraint is lifted nearly 200% of recent gasoline
demand could be met by liquid biofuels, corresponding to a scenario in which
all current cropland is converted to high- yielding switchgrass.
The size of our supply estimates indicate that while domestic biofuels can play
a large role in transportation, achieving such high levels of ethanol production
may not be socially or ecologically desirable, or may be extremely costly with
costs expressed through higher food prices, biodiversity loss, water degradation,
and soil erosion. Policies designed to protect natural resources and stabilize
food prices should be implemented early in order to achieve a reasonable level
of biofuel production that avoids pushing these boundaries.
Biofuel Boundaries:
Estimating the Medium- Term Supply Potential
of Domestic Biofuels
Andrew Jones, Michael O’Hare, Alexander Farrell
RESEARCH REPORT
UCB- ITS- TSRC- RR- 2007- 4
August 22, 2007
The Transportation Sustainability Research Center fosters research,
education, and outreach so that transportation can serve to improve
economic growth, environmental quality and equity. It is housed at
the UC Berkeley Institute of Transportation Studies.
http:// www. its. berkeley. edu/ sustainabilitycenter/
Biofuel Boundaries:
Estimating the Medium- Term Supply Potential of Domestic Biofuels
Andrew Jones, Michael O’Hare, Alexander Farrell
A UC Berkeley Transportation Sustainability Research Center
Working Paper
updated
August 22, 2007
2
Abstract
We estimate the physical supply potential of biofuels from domestic municipal solid waste,
forestry residues, crops residues and energy crops grown on existing cropland using optimistic
assumptions about near- term conversion technologies. It is technically feasible to produce a
significant amount of liquid biofuel ( equivalent to 30- 100% of 2003 gasoline demand) without
reducing domestically produced food and fiber crops or reducing the total calories available as
domestic animal feed. Most of this supply can be attributed to the potential of energy crops, with
the combination of municipal solid waste and forestry residues supplying between 10% and 30%
of recent gasoline demand.
Our modeling approach to energy crops is unique in that it explicitly models interactions
between the feed and fuel system using an optimization procedure that adjusts cropland
allocation among major crops subject to a simple food security constraint. Our modeling
indicates that sizable increases in biofuel production need not result in decreased availability of
food or animal feed, but will require changes in the composition of livestock diets away from hay
and soymeal toward either whole corn or feed coproducts of biofuel processing such as distillers
grains. Whole corn yields very high levels of digestible calories per land area, so shifting away
from soymeal and hay to corn feed permits the same total level of digestible calories to be
produced from a smaller area. Furthermore, the coproduction of animal feeds with biofuels
relaxes the need to grow dedicated feed crops at all. Thus, under our food security constraint,
energy crops which yield feed coproducts ( such as corn ethanol) can be grown on a larger area
than other energy crops, potentially yielding higher total levels of biofuel than other crops ( such
as switchgrass) that yield more biofuel but less animal feed per land area. When the food security
constraint is lifted nearly 200% of recent gasoline demand could be met by liquid biofuels,
corresponding to a scenario in which all current cropland is converted to high- yielding
switchgrass.
The size of our supply estimates indicate that while domestic biofuels can play a large role in
transportation, achieving such high levels of ethanol production may not be socially or
ecologically desirable, or may be extremely costly with costs expressed through higher food
prices, biodiversity loss, water degradation, and soil erosion. Policies designed to protect natural
resources and stabilize food prices should be implemented early in order to achieve a reasonable
level of biofuel production that avoids pushing these boundaries.
3
Table of Contents
Abbreviations Used..................................................................................................................... 4
List of Figures........................................................................................................................ .... 4
List of Tables......................................................................................................................... .... 4
Introduction ............................................................................................................................... 5
Methods........................................................................................................................ ............. 7
Feedstocks Considered............................................................................................................ 7
Wastes......................................................................................................................... ....... 7
Primary Resources .............................................................................................................. 9
Biofuel Conversion Processes Considered............................................................................. 16
Results ............................................................................................................................... ...... 18
Municipal Solid Waste .......................................................................................................... 18
Forestry Resources................................................................................................................ 20
Energy Crops and Crop Residues .......................................................................................... 22
Summary........................................................................................................................ ...... 25
Discussion ............................................................................................................................... 26
Appendix A - Biofuel Production Pathways .............................................................................. 28
Biofuel Feedstocks ............................................................................................................ 29
Biofuel Conversion Processes ........................................................................................... 30
Bibliography................................................................................................................... ......... 32
4
Abbreviations Used
MSW Municipal Solid Waste
CRP Conservation Reserve Program
DGS Distillers Grains with Solubles
DOE Department of Energy
USDA United States Department of Agriculture
List of Figures
Figure 1 Current Cropland and Rangeland Allocation ............................................................... 15
Figure 2 Feed and Fuel Yield of Selected Crops........................................................................ 15
Figure 3 Municipal Solid Waste Generation and Recovery........................................................ 18
Figure 4 Potential Ethanol Production from MSW Relative to Gasoline Consumption .............. 19
Figure 5 Forestry Related Feedstocks for Bioenergy ................................................................. 20
Figure 6 Potential Ethanol Supply from Forestry Related Feedstocks ........................................ 21
Figure 7 Gasoline Demand Met By Energy Crops by Scenario.................................................. 23
Figure 8 Feed Market Changes by Energy Crop Scenario.......................................................... 23
Figure 9 Detailed Land Use Change by Scenario....................................................................... 24
Figure 10 Summary of Biofuel Supply Potential Under a Food Security Constraint .................. 25
Figure 11 Biofuel Production Pathways .................................................................................... 28
List of Tables
Table 1 Assumed Crop Yields................................................................................................... 13
Table 2 Crop Scenarios ............................................................................................................. 14
Table 3 Biofuel Yield Per Dry Ton of Feedstock....................................................................... 17
Table 4 Estimated Cellulosic Feedstocks from MSW................................................................ 19
Table 5 Estimated Forestry Related Feedstocks......................................................................... 21
5
Introduction
As a distributed source of liquid transportation fuels, biomass may play an important role in
achieving medium- term energy security goals in the US. Furthermore, liquid biofuels can be
produced so that their life- cycle greenhouse gas emissions are significantly lower than those of
conventional transportation fuels, offering the possibility of reduced climate effects from the
transportation sector. However, liquid biofuels are unlikely to be a “ silver bullet” for solving
our energy security and climate change problems because of supply limitations. Just how much
of a role biofuels can play in meeting these twin objectives depends on several constraints.
A rational policy approach to biofuels will recognize these constraints and package biofuel
incentives with a portfolio of policies aimed at reducing the overall demand for transportation
fuels via vehicle efficiency and smart growth as well as policies promoting other low carbon
distributed fuels such as electricity. In addition, variation in biofuel production practices
warrants policies that address the life- cycle impacts of biofuels on food security, global climate,
regional climate, soil, air, water, and biodiversity. Biofuel policies may also need to distinguish
among biofuels according to their life- cycle liquid fuel consumption in order to maximize energy
security benefits. However, relatively little gasoline and diesel are typically used in the
production of biofuels ( 6). Thus the net supply of liquid fuels – the final product less the liquid
fuels used to make the product – often approaches the gross supply of liquid biofuels.
Physical constraints on biofuel supply include the availability of feedstocks and the efficiency of
converting feedstocks into fuels. For example, agriculture- derived feedstocks are constrained by
land, water, and nutrient resources as well as limitations on the efficiency of photosynthesis.
Many processes for converting biomass into liquid fuels involve external energy inputs, such as
electricity, process heat, or hydrogen in addition to the biomass itself. For any given process,
there is a maximum fraction of all input energies that can be preserved in the final fuel.
The cost of producing biofuels also limits supply. In well- functioning markets, resources that are
difficult to access or have valuable alternative uses will not be developed to their full physical
potential. However, markets for both energy and agriculture are severely distorted by subsidies,
mandates, and tariffs ( 10) Furthermore, these markets do not value several important social and
environmental externalities associated with biofuels. For instance, some biofuels may be
undervalued for their ability to mitigate global climate change, whereas the alternative use of
biomass for food may be undervalued to the extent that poor people cannot afford to pay for
what is actually of infinite value, their own subsistence. Furthermore, the price of biomass
generally does not reflect the unpriced costs of energy cropping on soil, water, biodiversity, and
climate. While such costs are the beyond the scope of the present analysis, they are vital
considerations. Developing biofuels within the bounds of ecological and human health will
require governance mechanisms to impose new constraints that present markets fail to impose.
In this study, we put considerations of cost and practicality aside to estimate the “ outer- limits”
physical supply potential of biofuels from US domestic resources using existing or near- term
technologies. The magnitude of this potential is of policy significance as an upper bound on
biofuels’ contribution to global warming ( or petroleum displacement) goals. We do this by
6
examining data on various biomass resources and conversion processes, and for the case of
energy crops, by developing scenarios of new crop distributions.
In most of our scenarios, we impose one external constraint – that the total amount of food, and
animal feed for domestic consumption, remains constant. We chose to consider this constraint
and not others because of its analytical tractability and high value to society. 1 However, for
completeness, we do present two scenarios in which food and feed production are completely
displaced by energy crops. In principle, an economy can forgo food production entirely and trade
for it, as long as the rest of the world is willing to increase production enough; for example,
Singapore, a city- state with trivial amounts of agricultural land, imports all its food and enjoys a
very high standard of living. For the United states to move significantly from being a food
exporter to being an importer of even a large fraction of its food would be a barely imaginable
earthquake on the world trade scene and impossible given domestic U. S. politics, but we model
this case on the “ upper bound” principle.
The ecological constraints mentioned above are serious considerations for a world with high
levels of biofuel development ( particularly from energy crops) and should not be taken lightly.
We discuss such constraints qualitatively throughout the text. Similarly, the internalized costs of
procuring diffuse resources and processing low- quality resources are considered only
qualitatively in this study. We do not consider direct logging of existing forests for bioenergy
feedstock. Due to the ecological sensitivity of such an endeavor, we believe that more advanced
scenario modeling would be required to evaluate tradeoffs among competing uses and values of
forestland. We do consider forest industry wastes and residues as discussed below.
We chose to focus on domestic resources because of the political imperative to develop such
resources for energy security and economic development purposes. Although US energy
security and climate benefits can be accomplished via imported biofuels, domestic biofuels have
two important strategic advantages. First, domestic resources are unlikely to suffer from supply
disruptions due to geo- political events, and second the greenhouse gas impacts of domestic
biofuel production can be readily monitored and regulated. On the other hand, it may be the case
that imported biofuels can be produced with more greenhouse gas efficient production pathways,
or that the magnitude and stability of international supply may be of interest from an energy
security perspective. These issues are not addressed here.
In the following sections, we first discuss the methodology used to estimate feedstock
availability and potential fuel supply from wastes and primary resources. Crop scenarios and
land use assumptions are discussed in this section. We then discuss conversion technologies.
Results are presented in the same order as methodology with wastes followed by primary
resources. The paper concludes with a short discussion and an appendix on conversion
technologies and fuels.
1 Note the crudeness of this analytical approach to food security. First, this approach does not
recognize the likely increase in food prices induced by competition for agricultural resources by
fuel crops. Second, it is conceivable that food security be met with less domestic production of
food and animal feed, particularly with a dietary shift away from grain- fed meat and dairy to
more resource efficient consumption, or through secure trade arrangements.
7
Methods
The goal of this study is to describe the envelope of upper bounds on resource availability,
deliberately extending the analysis into a range of tiny probabilities of occurrence. We consider
several categories of potential biomass feedstocks that have been discussed in the literature on
biofuels. For each supply category, we provide a low and high estimate. This range is driven by
uncertainty in biomass feedstock availability rather than uncertainty in conversion yields. The
biofuel yield per unit of feedstock used is the highest value for any well- documented conversion
technology available in the near- term ( 5- 10 years). See appendix A for an overview of biomass
feedstocks and conversion pathways, including near and long- term technologies. All of the
calculations were performed in a simple spreadsheet designed to allow the easy manipulation of
key assumptions such as conversion and crop yields with clear documentation of primary data
sources2.
Feedstocks Considered
Two kinds of biofuel feedstocks are considered in this study: waste streams and primary
resources.
Wastes
Waste streams, such as municipal solid waste and logging industry residues, are byproducts of
existing activities. Their development for ethanol may improve the economics of engaging in
those activities, but those activities are not likely to increase dramatically in order to supply
ethanol feedstocks. Thus the present study estimates the supply potential of waste feedstocks by
simply examining current industry trends. It should be noted that waste streams are not
necessarily “ free” in the sense of having no opportunity cost. Logging residues may play an
important role in soil retention and wildlife habitat if left in the forest. Similarly, a significant
amount of municipal solid waste is currently diverted from landfills for recycling, composting, or
combustion for energy.
Our low estimate of waste feedstock supply excludes wastes currently used for other economic
purposes, whereas the high estimate assumes that the entire wastestream is used for biofuel
production. Thus, in the high estimate calculations, we implicitly assume that price mechanisms
favor biofuel development over alternative uses without violating our definition of waste ( i. e.
that economic use of the feedstock does not encourage further feedstock production).
2 Please contact the author for access to this spreadsheet. It is our hope that it can easily be
adapted to incorporate new scenarios and better data as the relevant industries evolve.
8
Municipal Solid Waste ( MSW)
Municipal solid waste ( MSW) typically refers to the mixture of wastes collected in municipal
areas that is sent to municipal landfills, incinerators, or recycling facilities ( 2). It generally does
not include industrial process wastes, hazardous waste, construction and demolition debris,
automobile waste, sewage sludge or agricultural wastes, some of which are disposed of in
municipal landfills. We address the potential using construction and demolition debris and some
agricultural wastes elsewhere in this report.
Barriers to biofuel generation from MSW include the ability to effectively separate cellulosic
material from other wastes, potential variation in feedstock quality and availability, as well as the
cost of handling and drying very moist materials and competing uses for MSW such as recycling,
compost, and landfill gas generation. The use of MSW for biofuel -- when economically
practical -- has very few drawbacks and even has the positive effect of reducing demand for
landfill space. However, MSW is a noxious material and potentially hazardous, making the siting
of production facilities challenging. From a climate perspective, fuel derived from MSW will
likely have low life- cycle greenhouse gas emissions because upstream emissions – those
involved in generation of the feedstock – would have been generated regardless of the use of
MSW. On the other hand, the alternative fate of carbon in landfills matters too. If landfilled
MSW is sequestered, its diversion to biofuel production represents a gain of carbon in the
atmosphere, whereas if landfilled MSW leads to the release of methane, its diversion may
represent a decrease in climate forcing because methane is more potent as a greenhouse gas than
carbon dioxide.
We considered two studies of MSW generation in the US, which represent a range of uncertainty
about actual MSW generation rates ( 2, 21). The first report that we considered is from the EPA
and is based on a mass balance methodology3 ( 2). This report estimates that 246 million tons of
MSW were generated in 2005. The methodology in this report also permits categorizing the
waste stream in terms of product source and material. Of the 246 million tons generated in 2005,
160 million tons were potential cellulosic biofuel feedstocks in the form of paper and
paperboard, yard trimmings, food waste, wood, and other organics. However, some of this
material ( 64 million tons) was recovered for recycling and compost. After waste recovery, 95
million tons of potential cellulosic feedstock was landfilled or burned ( 2). Plastics and textiles
are also potential feedstocks for fuels derived through non- enzymatic conversion pathways, such
as gasification. However, the current study considers only the enzymatic pathway as discussed
in the methods section and so plastics are not considered.
The second report that we considered used surveys to obtain information directly from each state
about MSW generation, recycling, combustion, and landfilling ( 21). Different states use
different methods to track MSW including direct sampling of the wastestream, thus the accuracy
of the report depends upon the accuracy of individual state data. The report aggregated data
from each state, adjusting for inconsistencies in the definition of MSW. This report estimated
that MSW generation in 2004 was 388 million tons, nearly 60% higher than the EPA report.
3 This method applies adjustment factors that have been developed by the EPA to data on
products imported, exported, and produced to estimate the generation of different types of MSW.
9
However, the survey methodology did not permit categorizing the material composition of
MSW.
Rather than attempt to assess the accuracy of either of these two reports, we accept their
variation as an uncertainty range on the availability of MSW. Because the second report does
not consider the material composition of MSW, we apply the relative fraction of MSW
composition from the first study to the second
Forestry Residues
The forest products industry generates residues and wastes at every level of production from
logging to the disposal of construction debris. We considered several categories of such
materials, as estimated by the USDA Forest Service in the DOE Billion Ton Vision ( 18). These
include logging residues - the upper portions of trees currently left behind when an area is
logged, forest products industrial residues - such as black liquor and sawdust, and construction
and demolition debris. We also considered the potential availability of forest thinnings, which
are small diameter trees and undergrowth removed from managed forests in order to reduce the
risk of catastrophic fire. The presence of such material is a major problem in many forests due to
historically misguided management. The current cost of thinning forests is often prohibitive, and
use of thinned material for biofuel conversion may improve the economics of fire management.
Each of the resources listed above may have significant costs associated with their use as biofuel
feedstocks. Logging residues play an important ecological role in preventing soil erosion,
nutrient cycling and providing wildlife habitat, in addition to being a diffuse resource that may
require specialized equipment to harvest and collect for biofuel conversion. Forest industry
residues such as black liquor are mostly ( 93%) ( 18) already used to generate energy on- site in
facilities such as paper mills. Diversion of these resources to biofuel production will require
substituting natural gas, electricity, or even coal to provide industrial process energy.
Construction and demolition debris, like MSW may vary in quality and may be costly to collect
to a central facility. Finally, forest thinnings are very diffuse and costly to collect.
In the Billion Ton Vision report, USDA provides an estimate of additional residues that might
become available as the forest products industry expands and becomes more efficient. This
estimate includes residues from each of the categories discussed above and is based on a
technical assessment of the US forestry industry from 1952- 2050 ( 8). We take this projection to
be speculative and so include it only in our high estimate of feedstock availability.
Primary Resources
Primary resources, such as land and water, can be used to produce energy crops for biofuel
production. The potential supply of biofuels from primary resources is more complicated to
estimate than the supply from wastes. This is because the feedstock supply itself can’t be taken
as a fixed complement of existing activities, but results from the re- allocation of resources away
from current uses. For example, the existing corn crop has been partially diverted away from
10
feed and export use toward ethanol production and more land has been brought into corn
production, increasing the total amount of corn and reducing the supply of soy and other crops.
Alternatively high- yielding switchgrass could be grown as a cellulosic ethanol feedstock instead
of existing crops, but processing switchgrass into ethanol does not create the animal feed
coproduct that corn ethanol does, nor does it have the option value of being sold in grain markets
if the ethanol market is unstable. We do not attempt to weigh these tradeoffs here. Rather, we
explore the technical limits within which markets and policy can shape outcomes.
Crop residues occupy a grey area between waste stream and primary resource. This is because
the ability to collect and use crop residues for biofuels may actually influence the development
of certain crops over others. For instance, the ability to collect and sell corn stover, the stalks
and leaves of the corn plant, as a biofuel feedstock may bias the production of corn over
switchgrass. In this study crop residues are treated in conjunction with energy crops and
represent an increase in the effective biofuel yield per unit of land area from corn and other
crops.
Energy Crops and Crop Residues
Agricultural lands supply food for domestic consumption, large quantities of animal feed, fiber
for textiles, and export commodities. Figure 1 shows the current allocation of US cropland to
major crops. Grassland pasture and range, which is typically so dry and unproductive of biomass
that its only economic use is for foraging animals, is included for reference. While the
possibility to grow bioenergy crops on rangeland and pastureland does exist, this study only
considers crops grown on prime cropland ( including conservation reserve program land). This is
because yield data for corn and switchgrass, the two model energy crops considered are based on
assumptions of adequate soil and water resources.
Of the 440 million acres of land classified as cropland by the USDA, approximately 60 million
acres is used for cropland pasture, and 40 million acres is classified as idle cropland, which
includes the Conservation Reserve Program ( CRP) lands4. Of the remaining 340 million acres,
260 million acres are dedicated to four major crops: soybeans ( 74 million acres), corn ( 74 million
acres), hay ( 62 million acres), and wheat ( 50 million acres). The remaining 80 million acres are
used to grow cotton, sorghum, small grains, oilseeds, fruits, vegetables, nuts, and tobacco. All
fruits vegetables and nuts are grown on just 8 million acres. ( 1)
Approximately 210 million acres ( 47%) of cropland is used to produce animal feed or used for
cropland pasture ( see Figure 1) ( 1), that is, for meat and milk. Cropland characterized as pasture
may be marginal or degraded cropland not suitable for high yielding energy crop production.
Thus, cropland pasture is excluded from the present analysis. This leaves 150 million acres
( 34%) of land used exclusively to produce crops for animal feed, mostly corn, soybeans, and
4 The CRP is a USDA’s primary land retirement program which pays farmers to take
environmentally sensitive land out of production. Many of these are productive cropland that
happen to have highly erodable soils.
11
hay. When we refer to feed crops in the scenarios, we mean crops grown on these 150 million
acres. Approximately 60 million acres ( 14%) of cropland is used to produce exports. This area
is composed mostly of soy, corn, wheat, and cotton ( 1).
We estimate the potential role of energy crops and crop residues as biofuel feedstocks by
developing several scenarios. Each scenario alters the existing allocation of cropland to include
more bioenergy crops, assuming that no new land is brought into crop production and that the
yields of various crops are constant across the land area currently used for crops. Thus existing
pasture, rangeland, forest and other non- crop uses remain unchanged in each scenario, but the
portion of cropland used for soy, hay, or other existing crops may be reduced to accommodate
more energy crops.
We chose to examine two model energy crops: corn and switchgrass. Corn forms the basis of the
existing biofuels industry and is likely to continue to play an important role for the foreseeable
future. Producing ethanol from corn has the added benefit of yielding a protein- rich animal feed
coproduct in the form of distillers grains ( DGS) or similar products such as corn gluten meal.
Because our analysis is focused on the near term, we assume a recent corn crop yield of 160
bushels per acre.
Additional biofuels can be produced from an acre of corn by harvesting the cellulosic residue
known as corn stover, which we model in some of our scenarios. Because the removal of stover
from corn fields negatively affects soil erosion and soil quality, considerable debate exists about
the appropriate level of removal ( 13, 16). Consistent with the upper- bound approach, we
assume that a high fraction ( 75%) of corn stover is harvested in those scenarios. In practice,
acceptable removal rates will depend on regional yields, climatic conditions, and cultivation
practices ( 23).
Switchgrass is a high- yielding native perennial grass that has received much attention as a
prospective energy crop( 15). Switchgrass and other herbaceous perennial crops are not
necessarily associated with multiple uses and coproducts like corn. However, the potential yield
per acre of dedicated energy crops such as switchgrass are higher than for corn and the perennial
nature of the crop allows for the possibility of lower inputs, decreased soil erosion, and increased
soil carbon accumulation compared to annual crops such as corn( 11, 14). For these reasons, and
because switchgrass can only be utilized as a biofuel feedstock using more advanced conversion
technologies, the expected life- cycle greenhouse gas emissions associated with switchgrass
ethanol are much lower than for corn ( 6). In our model, we assume that breeding and cultivation
improvements will increase the current average switchgrass yield of 13.5 Mg/ ha to the currently
highest reported yields of 22 Mg/ ha( 15).
As a result of crop re- allocation, the amount of export, food, and animal feed production changes
in our scenarios. All but two scenarios utilize a simple food security constraint that holds all
cropland used for domestic food production constant and holds the amount of calories available
for animal feed constant while a linear optimization algorithm maximizes biofuel production.
For instance, distillers grains, a protein- rich coproduct of corn ethanol production may comprise
a larger fraction of animal feed diets in these scenarios, but the same or greater magnitude of
12
digestible calories are available to livestock despite diversion of corn grain from the feed system.
For simplicity, digestible calories are based on beef and dairy net energy values( 4, 19) and
expressed in tons of corn equivalent. Cattle are ruminants and can more easily utilize the energy
in many feeds compared to swine and poultry, so our digestible calorie factors may be
overestimates from the perspective of the average feed consuming unit. However, cattle
consume a large portion of the feed currently consumed ( 9).
We did not consider the availability of protein as animal feed constraint. Ignoring protein
availability is justifiable from the standpoint of converting existing feed corn into corn ethanol
because the protein value of the corn is preserved in ethanol coproducts such as distillers grains.
However, the conversion of soy acres to corn ethanol or switchgrass production may lead to a
deficit in protein available as feed. This would be an interesting area in which to expand the
detail of the model.
For most crops, determining the proportion of harvested land used for animal feed and exports
was simple because the harvested portion of the crop is itself the commodity that is traded or
used as feed ( e. g. corn, cotton, hay) ( 1). However, soybeans are often separated into two
different products – soymeal and soybean oil – which are not exported or used for feed in the
same proportions. We chose to track the fate of the soymeal in determining whether a particular
unit of soybeans was used for feed, exports, or other uses. Tracking the fate of soybean oil
would underestimate the use of soy as animal feed. However, this choice partially undermines
the food security constraint because scenarios in which soybean production declines may be
associated with decreases in domestic soybean oil production.
The crop scenarios combine information from different time periods in order to use the most
recent data available. For instance, data on gross allocation of land to cropland, rangeland,
pasture, etc. was based on USDA’s 2002 Census of Agriculture( 12), which is not available for
more recent years, whereas allocation of cropland to specific crops such as corn for grain, corn
for ethanol, soybeans etc. is based on more recent data from 2004( 1). We chose to use 2004 data
to reflect recent trends in corn production related to the development of corn ethanol. In general,
we used the most recent data available for a given parameter or we selected hypothetical
assumptions about future conditions from the literature.
Crop yield assumptions were taken from a variety of sources and are summarized in Table 1 ( 1,
15, 16, 22). Figure 2 combines crop yield data with assumptions about biofuel conversion
processes and animal feed values to illustrate the feed and fuel yields of selected crops in our
model. Protein yields are included for reference, although they do not play an essential role in
our model at this time. It should be noted that although high yielding switchgrass produces more
fuel energy per unit of land than corn ethanol ( even with stover collection), corn ethanol also
produces animal feed coproducts. These tradeoffs are important drivers of the modeling results.
13
Table 1 Assumed Crop Yields
Crop yield assumptions were taken from a variety of sources as listed here.
Assumed Crop Yields
Dry Mg /
Harvested Ha Source
Corn Grain 10.1 NASS 2005 Ag Statistics( 1)
Corn Stover ( at 75% harvest rate) 7.5 Nelson et al 2004( 16)
Soy Bean 2.9 NASS 2005 Ag Statistics( 1)
Switchgrass ( current average) 13.5
Argonne Lab GREET
model( 22)
Switchgrass ( high yield) 22.0 Mclaughlin et al, 2005( 15)
Crop Scenarios
We considered 4 corn ethanol scenarios and 2 switchgrass scenarios. Information about each
scenario is summarized in Table 2 and discussed in more detail below. Two scenarios – C4 and
S2 consider the extreme case in which all cropland is used to produce either corn or switchgrass
for biofuel. The other scenarios employ a simple food security constraint which optimizes
biofuel production on the land area permissible for biofuel production in each particular scenario
while maintaining domestic food production ( fruits, vegetables, nuts, as well as wheat, soy, and
corn used directly for food or seed) fiber production ( cotton for domestic use) and a constant
level of feed calories for domestic livestock as discussed above.
14
Table 2 Crop Scenarios
We considered four corn scenarios and two switchgrass scenarios grown either on land currently used for feed
crops, land currently used for feed and export crops, or on all cropland. High- yielding corn refers to corn with 75%
stover collection and high- yielding switchgrass refers to the highest reported yields at this time. The final column
refers to whether or not the food security constraint was applied as discussed in the text.
In scenario C1, we considered corn ethanol production on land currently used to produce
domestic feed crops only. The corn yield is based on recent years, and the food security
constraint applies.
Scenario C2 is like scenario C1, except that land used to produce exports is used to produce corn
ethanol as well.
Scenario C3 considers the collection of corn stover for ethanol production in addition to the corn
grain. This effectively results in a higher yield of ethanol per unit of land area dedicated to corn
ethanol production, as well as an additional ethanol yield from corn grown for other purposes.
The food security constraint still applies.
In scenario C4, the food security constraint is lifted and we compute the maximum physical
potential of corn ethanol production with stover removal on all existing cropland using near term
technologies.
Scenario S1 considers high yielding switchgrass on land currently used for feed and export crop
production with a food security constraint.
Scenario S2 is like scenario C4 in that the food security constraint is lifted to compute the
maximum technical potential of ethanol from switchgrass grown on all existing cropland.
Corn Scenarios
Switchgrass
Scenarios
Potential Land Crop Yield Food Constraint?
C1 Feed Crops current yes
C2
Feed and Export
Crops
current yes
C3 S1
Feed and Export
Crops
high yes
C4 S2 All Cropland high no
15
Figure 1 Current Cropland and Rangeland Allocation
Of the 440 million acres of land classified as cropland by the USDA, approximately 60 million acres is used for
cropland pasture, and 40 million acres is classified as idle cropland, which includes the Conservation Reserve
Program ( CRP) lands. Of the remaining 340 million acres, 260 million acres are dedicated to four major crops:
soybeans ( 74 million acres), corn ( 74 million acres), hay ( 62 million acres), and wheat ( 50 million acres).
Approximately 150 million acres is dedicated to crops grown for animal feed and an additional 50 million acres is
for growing export commodities. In 2004, only 11 million acres were used for biofuel production. For comparison,
590 million acres are considered grassland pasture and range,
Figure 2 Feed and Fuel Yield of Selected Crops
According to our crop yield and process yield assumptions, corn grain yields high levels of animal feed calories per
acre ( expressed as tons of corn equivalent). About 1/ 3 of the feed calories and all of the crude protein in corn are
available as distillers grains when corn is converted to ethanol, which is still more calories per acre than soybeans.
Although high- yielding switchgrass produces more ethanol per acre than corn with high levels of stover collection,
we do not assume that switchgrass produces a feed coproduct. Soybeans produce more protein per acre than corn.
Prairie is included for reference, assuming aboveground biomass can be converted using cellulosic
saccharification.
16
Biofuel Conversion Processes Considered
For each feedstock considered we chose the best- yielding near- term conversion process for
which reasonable yield data was available. Consequently, two conversion processes were chosen
from among the many possibilities outline in Appendix A: the conventional drymill starch- to-ethanol
process widely used today and a cellulose- to- ethanol process. 5 The assumed yields per
dry ton of feedstocks are given in Table 3.
The starch- to- ethanol process was applied to corn grain feedstocks and yield is based on the
latest generation of drymills in use in 2005( 22). The cellulose- to- ethanol process was applied to
all other potential feedstocks with the exception of switchgrass and is based on the Department
of Energy’s Theoretical Ethanol Yield Calculator( 3). This calculator provides a theoretical
maximum yield given the composition of various sugars in the feedstock. According to the
DOE, the expected practical yield is in the range of .6 to .9 times the theoretical maximal yield.
In the spirit of creating upper- bound estimates, we chose a factor of .9 for all cellulosic yields.
Thus variation in our final biofuel supply estimates reflects uncertainty in feedstock availability
rather than process yields. Switchgrass cellulose- to- ethanol yields were based on the GREET
model( 22).
Because the DOE calculator is based on dry tons of feedstock, we were forced to make
assumptions about moisture content of feedstocks in cases where this information was not
available. Similarly, when information about the proportion of various sugars was not available,
we matched unknown feedstocks with similar feedstocks for which sugar composition is well
documented. Theoretical maximum ethanol yield varied from 82 gallons of ethanol per dry ton
for forest thinnings to 116 gallons of ethanol per dry ton for mixed paper.
5 For some heterogeneous cellulosic feedstocks, such as MSW, the gasification to Fischer-
Tropsch fuel pathway may turn out to be a more appropriate pathway due to its relative tolerance
of feedstock variation. However, at this time, reliable yield data is not available because of the
paucity of research on biomass gasification to FT- fuels compared to enzymatic cellulosic
ethanol.
17
Table 3 Biofuel Yield Per Dry Ton of Feedstock
This table lists the conversion process, fuel produced ( ethanol in all cases), and assumed yield for each feedstock
considered in this study.
Feedstock Conversion Process
Yield ( GGE
/ Dry Ton)
MSW
Paper and Paperboard 70
Wood 61
Food, Other Organic 70
Yard Trimmings
cellulosic saccharification
49
Forestry
Logging Industry Residues 49
Fuel Treatments ( Forest Thinnings) 49
Forest Products Industry Residues 61
Construction and Demolition Debris
cellulosic saccharification
61
Energy Crops
Corn dry mill fermentation 72
Switchgrass cellulosic saccharification 67
Crop Residues
Corn Stover cellulosic saccharification 68
18
Results
Municipal Solid Waste
Figure 3 demonstrates the range of MSW generation estimates and relative breakdown of
different materials. Brackets indicate the portion of the MSW stream that represents potential
cellulosic ethanol feedstocks.
Our high estimate of MSW feedstock availability is based on the higher estimate of total MSW
generation and assumes that all potential cellulosic feedstocks are used for cellulosic ethanol
production, including feedstocks that are currently recycled or combusted for energy. The low
estimate uses the lower estimate of total MSW and assumes that existing recycling and recovery
program remain intact. These values are summarized in Table 4, along with our assumed
moisture content for these feedstocks.
Based on the cellulosic ethanol assumptions discussed in the methods section, we estimated the
technical potential of supplying ethanol from MSW. These results are presented in comparison
to recent gasoline demand in Figure 4. At current rates of waste generation, MSW has the
technical potential to supply between 2 and 9 percent of the 2003 US gasoline demand, mostly
from paper and paperboard feedstocks. The cost of collecting and handling MSW feedstocks
may be a significant cost barrier to developing this level of biofuels from MSW as discussed
above. A significant fraction ( about 50%) of current waste paper and paperboard is separated
from the wastestream for recycling, so the infrastructure for this separation already exists.
However, society must weigh the value of biofuel against the opportunity cost of not recycling
these materials.
Figure 3 Municipal Solid Waste Generation and Recovery
There is a significant discrepancy in the literature regarding how much MSW is generated in the US ( represented by
the first two columns). Portions marked with brackets represent potential cellulosic feedstocks. The third column
represents the amount of MSW that is not currently recycled, composted, or combusted for energy.
19
Table 4 Estimated Cellulosic Feedstocks from MSW
This table lists our assumptions about the availability of MSW feedstocks. The high and low estimates reflect
variation in published studies of MSW generation. Furthermore, the low estimate excludes materials currently
utilized for recycling, composting, and combustion.
High
Estimate
( million
tons)
Low
Estimate
( million
tons)
Assumed
Moisture
Content
Paper and Paperboard 133 28 10%
Wood 22 10 25%
Food, Other Organic 46 22 75%
Yard Trimmings 51 7 50%
Total 251 67
Figure 4 Potential Ethanol Production from MSW Relative to Gasoline Consumption
Biofuels derived from MSW have the technical potential to supply between 2% and 9% of recent gasoline demand.
Most of this is from paper and paperboard, half of which is already recycled for non- energy uses.
2% 9%
20
Forestry Resources
The estimated yearly technical supply of forestry related wastes and residues is presented in
Figure 5 and summarized in Table 5. The two differences between the high and low estimates
are that the high estimate includes an estimate of additional residues from an expanded industry
( as discussed in the methods section) and the low estimate excludes materials currently used for
energy services. These feedstocks have the technical potential to supply 5 % to 14% of recent US
gasoline demand. These results are presented in Figure 6.
Figure 5 Forestry Related Feedstocks for Bioenergy
Between 130 and 320 million tons of forestry related feedstocks are available or might become available for biofuel
conversion. The high estimate includes an estimate of additional residues from an expanded industry ( see methods)
and the low estimate excludes materials currently used for energy.
21
Table 5 Estimated Forestry Related Feedstocks
Between 130 and 320 million tons of forestry related feedstocks are available or might become available for biofuel
conversion. The high estimate includes an estimate of additional residues from an expanded industry ( see methods)
and the low estimate excludes materials currently used for energy.
Material
High
Estimate
( Million
Dry Mg)
Low
Estimate
( Million
Dry Mg)
Logging Industry Residues 41 41
Fuel Treatments ( Forest Thinnings) 60 60
Forest Products Industry Residues 116 8
Construction and Demolition Debris 20 20
Additional Residues from Expanded Industry 89 0
Total 326 129
Figure 6 Potential Ethanol Supply from Forestry Related Feedstocks
Between 5% and 14% of recent gasoline demand could technically be met by forestry related feedstocks. The higher
estimate would require diversion of forest products industry residues, such as black liquor from their current use for
industrial process energy. Furthermore, the high estimate includes an assumption that the entire forest products
industry will expand due to economic and population growth.
5% 14%
22
Energy Crops and Crop Residues
Figure 7 shows the amount of transportation fuel produced in each of the scenarios that we
modeled. The only scenarios in which enough fuel is produced to completely meet our current
( 2003) gasoline demand are the extreme scenarios in which all cropland is used to produce either
corn ethanol or switchgrass for ethanol. In the more modest scenarios in which food security
constraints are applied, between 25% and 76% of recent gasoline demand is met with biofuels. If
all cropland were dedicated to either corn or switchgrass, 135% or 172% of recent gasoline
demand could be met by biofuels, respectively.
The amount of fuel produced in a given scenario is determined by the parameters of the scenario
( the energy crop considered and the land area suitable for energy crop production) as well as the
interaction between the feed and fuel sectors which results from the food security optimization.
Figure 8 shows the mix of livestock feeds produced in each scenario. As expected, a constant
level of feed calories ( expressed as million ton corn equivalents) is produced in those scenarios
with food security constraints. In general, the baseline feed mix is shifted either towards more
whole corn grain or toward corn ethanol coproducts ( DGS) and away from roughage ( hay) and
soymeal. Because corn yields higher levels of feed energy per unit of land than other feed crops
such as soybeans ( see Figure 2), the optimization dynamics tend to produce more corn for feed in
order to free up land for energy crop production ( scenarios C1 and S1). This tendency is
moderated by the production of corn ethanol coproducts, however, so much so that in some
scenarios ( C2 and C3) whole corn feed is reduced due to the abundance of DGS. It should be
noted that large shifts toward corn production will be accompanied by increases in the
externalities associated with corn production such as nitrogen runoff and soil degradation.
By reducing the need to grow whole corn grain for feed, the production of DGS permits the
optimization procedure to grow corn for ethanol on a larger land area than switchgrass, which
does not produce a feed coproduct. This explains why scenario C3 produces more transportation
fuel than scenario S1, despite the fact that switchgrass produces more fuel per unit of land area.
When the food security constraint is lifted ( scenarios C4 and S2), switchgrass produces more
fuel than corn ethanol. However, scenario S2 is associated with a severe deficit in feed, whereas
in scenario C4 there is a glut of DGS.
Shifts in crop acreage associated with each scenario are presented in Figure 9. The largest shifts
are generally away from low feed calorie yielding crops such as hay and soy and toward energy
crops or corn for feed as discussed above. A small amount of wheat, cotton, and other crops are
reduced in those scenarios, which permit energy crops to be grown on land currently used for
exports. In scenarios C4 and S2, all cropland is dedicated to energy crops.
23
Figure 7 Gasoline Demand Met By Energy Crops by Scenario
In the presence of a food security constraint ( scenarios C1, C2, C3, S1), biofuels from energy crops can supply
between 25% and 76% of recent gasoline demand. If the constraint is lifted and all cropland is dedicated to biofuel
production, between 135% and 172% of recent gasoline demand could be met with biofuels.
Figure 8 Feed Market Changes by Energy Crop Scenario
In the presence of a food security constraint ( scenarios C1, C2, C3, S1), total feed calories ( expressed in tons of
corn equivalents) are held constant. In general roughage ( hay), soymeal, and other feeds are replaced by either
more whole corn grain or biofuel feed coproducts such as distillers grains. When all cropland is dedicated to corn
ethanol production, there is a glut of distillers grains, whereas when all cropland is dedicated to switchgrass, there
is an animal feed deficit.
25%
46%
76%
60%
172%
135%
24
Figure 9 Detailed Land Use Change by Scenario
In the presence of a food security constraint ( scenarios C1, C2, C3, S1), cropland tends to shift from the production
of low- yielding feed crops such as hay and soybeans toward more whole corn for feed or biofuel crops. In scenarios
C2 and C3, the high level of corn ethanol production eliminates the need for additional corn for grain due to the
reduction of distillers grains. In scenarios C4 and S2, all cropland is shifted to corn ethanol and switchgrass
production, respectively
25
Summary
Subject to a simple food security constraint, we estimate that the technical potential of supplying
transportation fuel using near- term conversion technologies is between 30% and 99% of 2003
gasoline demand ( see Figure 10). Variation between the high and low estimate is based on
uncertainty about feedstock supply rather than conversion technology yields, for which we use
optimistic assumptions. While MSW and forestry residues can supply a modest amount of
transportation fuel, the bulk of potential supply is from energy crops and crop residues. If the
food security constraint is lifted, then the potential supply from all sources nearly doubles to
195% of 2003 gasoline supply. This value is based on a scenario in which all current cropland is
used to grow high- yielding switchgrass for biofuels.
Figure 10 Summary of Biofuel Supply Potential Under a Food Security Constraint
Subject to a simple food security constraint, the technical potential of supplying transportation fuel using optimistic
assumptions about near- term conversion technologies is between 30% and 99% of 2003 gasoline demand.
Variation between the high and low estimate reflects uncertainty about feedstock supply rather than conversion
technologies.
30% 99%
26
Discussion
In this study, we estimate the physical supply potential of liquid biofuels from domestic
resources using near- term technologies. We estimate that nearly twice our current level of
gasoline consumption could be met by biofuels if we were willing to forgo all other crop
production. However, with a simple food security constraint, this value drops to 33%- 99% of
current gasoline consumption. Even these values reflect optimistic conversion yields and ignore
cost constraints – both economically valued costs such as the cost of harvesting diffuse resources
and external costs such as damage to ecological systems and human health. In particular, we did
not explicitly model the climate impacts of the various biofuel pathways considered in this study.
Limiting the analysis only to low- carbon biofuels may result in even lower supply estimates.
While several prominent studies have addressed the issue of domestic supply of biofuels relative
to fossil energy consumption for transportation( 7, 17, 18), ours is unique in that it explicitly
models interactions between the feed and fuel system using an optimization procedure that
adjusts cropland allocation among major crops. Our modeling shows that sizable increases in
biofuel production need not result in decreased availability of animal feed, but will require
changes in the composition of livestock diets away from hay and soymeal toward either whole
corn or feed coproducts of biofuel processing such as DGS. In our model, these changes in feed
composition are associated with land cover changes that favor more corn production. However,
the possibility exists for new bioenergy crops such as switchgrass to produce feed coproducts as
well( 7), indicating that this result may be more general.
The manner in which we model the feed- fuel interaction as a constrained optimization problem is
far cry from realistic. By fixing the level of food and feed consumption, we implicitly assume
that diets will not shift as increased fuel production drives up the price of food and feed. We
have also optimistically assumed that various feeds are perfectly substitutable on a digestible
calorie basis. In such a world, the value of the corn ethanol coproduct outweighs the additional
ethanol yield that switchgrass has compared to corn with stover collection. This is because the
corn ethanol feed coproduct relaxes the constraint on feed production and permits more land area
to be dedicated to fuel production.
Historically, transportation fuel demand has been rather unresponsive to price changes as well( 5,
20), which we have not modeled. Thus, it is entirely possible for fuel markets to bid up the price
of biofuel such that high- yielding switchgrass will be favored over corn ethanol, reducing the
supply of animal feed. However, the degree to which this happens will depend upon the price of
other substitutes for conventional petroleum and the price of petroleum itself. In theory, once
agricultural markets are open to high levels of biofuel production, the relative ability of
consumers to adjust levels of domestic food and fuel consumption ( either by decreased total
consumption, imports, or non- agricultural substitutes) will determine the level of food versus
fuel production, with the least flexible commodity tending to dominate. In any case, we have
demonstrated that it is technically feasible to maintain current levels of feed production while
significantly increasing biofuel production. However, this is no guarantee that the price of feed
and food products will not increase significantly. Food prices and the affordability of food are
27
complex issues with many stakeholders including domestic consumers, overseas farmers, and
food processors and retailers. Modestly higher food prices may be compensated for by dietary
shifts away from grain- intensive meat consumption or reduced consumption of other goods.
However, if food prices become too high, this may render our energy crop scenarios socially
unacceptable. In such a case, policy interventions may be needed to help food and feed crops
compete with bioenergy crops while maintaining reasonable food prices.
Waste resources are often overlooked in the discussion of biofuels. For instance, MSW is
excluded from the USDA and DOE’s Billion Ton Vision report ( 18), NRDC’s Growing Energy
report( 7) and from paper’s presented by biofuel skeptics ( e. g. ( 17)). We found that MSW can
probably play only a small role in supplying transportation fuel ( 2%- 9%) of recent gasoline
demand. However, these resources already have specialized collection and handling systems in
place and may come at negative cost due to avoided landfill fees. Furthermore, the climate
impacts ( and other external impacts) of converting waste to fuel are likely low compared to crop-based
biofuels. However, diverting wastes such as paper from the recycling stream may create
pressure on primary resource extraction.
As with MSW, forestry residues offer a modestly sized, but potentially uncontroversial and low
impact feedstock for biofuel production. The exception to this is logging residues, which do play
an important role in maintaining soil health and wildlife habitat. Diverting forest products
industry residues that are currently used for industrial process energy to liquid fuel production
will require new sources of process energy, which may result in additional greenhouse gas
emissions from fossil fuels. However, many more options exist for low greenhouse gas process
heat ( solar, nuclear, combined heat and power) than for low greenhouse gas liquid fuels. Thus
liquid fuel production might be considered a more desirable use of such feedstocks.
We have shown that while domestic biofuels may play a significant role in replacing or
supplementing conventional transportation fuels, they are by no means a “ silver bullet” that can
supply all of our needs. To replace all US gasoline consumption with biofuels, even under the
most optimistic yield assumptions, it would be necessary to harvest all the source material in
forestry residue and solid waste, eliminate agricultural exports, and either significantly shift
animal feed diets toward whole corn and corn ethanol coproducts or eliminate some degree of
domestic food and feed production. More realistically, biofuels should be thought of as one of
several items in a portfolio of strategies for lowering the climate impact and increasing the
security of our transportation system. Other strategies include smart growth, fuel efficiency, and
plug- in vehicle technology. To ensure the climate mitigation benefits of biofuels, green labeling
standards or a low- carbon fuel standard may be necessary. Research is needed to improve
conversion technologies, agronomic practices, and feedstock handling in order to accommodate
diffuse and heterogeneous feedstocks, lower costs, and minimize negative effects on climate,
ecosystems, and human health
28
Appendix A - Biofuel Production Pathways
“ Biofuels” describe transportation fuels derived primarily from ( recently grown, as opposed to
fossil) biological materials. There are several types of fuel that can be potentially produced from
biomass, multiple processing strategies to convert biomass to these fuels, and an immense range
of biomass feedstocks that could be utilized for one or more of these processes and fuels. Each
unique feedstock, conversion, and fuel combination is referred to as a fuel “ pathway.” The net
greenhouse gas emissions associated with a particular biofuel depends upon the entire fuel
pathway, and can vary greatly even among pathways for which the final fuels produced are
indistinguishable. Figure 11 provides an overview of selected biofuel production pathways that
are discussed in more detail below.
Figure 11 Biofuel Production Pathways
Feedstocks Processing Co- Products Fuels
Starches and Sugars Ethanol
Sugarcane Fermentation
Corn
Sugar Beet Butanol
Sweet Sorghum
Waste Sugars
Ligno- Cellulosic Crops
Switchgrass Animal Feed
High Diversity Grasses Sacharification
Poplar Electricity
Ligno Cellulosic Residues
Corn Stover
Rice Straw Gasification
Forest Residues, Thinnings
Orchard Prunings
Ligno- Cellulosic Wastes Methanol
Food Waste Catalysis
Yard Waste
Paper Waste Dimethyl
Other Municipal Solid Waste Ether
Construction and Demo Debris Flash
Forest Products Industry Waste Pyrolysis Fischer Tropsch
Fuels
Oils
Soy Hydrogenation Renewable
Rapeseed Diesel
Sunflower
Oil Palm Trans Biodiesel
Algae Esterification
Waste Oils
Hydrogen
H2 Producing Algae
29
Biofuel Feedstocks
Sugar and starch crops
Sugar crops, including sugarcane, sugar beets, and sweet sorghum, require relatively little
processing to derive the simple sugar sucrose for fermentation to alcohol by yeasts. Starch crops
such as corn, milo, or wheat require hydrolytic and enzymatic action to covert glucose and
fructose to sucrose.
Ligno- Cellulosic crops
The cell walls of plants are composed mostly of lignin and cellulose. Cellulose is a polymer
composed of starches that can be broken down into simpler components enzymatically through a
process known as saccharification. Both lignin and cellulose release thermal energy when
combusted for process heat, or when they undergo gasification or pyrolysis. Ligno- cellulosic
crops, both herbaceous and woody plants, represent a potentially more widely available biofuel
feedstock than sugar and starch crops Both herbaceous and woody crops are perennial, and
where they replace annual crops they are likely to increase soil organic carbon, creating a carbon
sink6. These crops may also have relatively low fertilizer and other input requirements, resulting
in a relatively low GHG profile. Furthermore, because ligno- cellulosic conversion processes
typically use the entire plant biomass either as direct feedstock or for process heat, the potential
yields per land area are generally higher than for agricultural crops.
Ligno- Cellulosic Residues
Residues may be collected as a by- product of the production of other crops, such as corn stover
or rice or wheat straw, or they may be collected after processing of other crops, such as lumber
mill, cotton gin, or vegetable processing residues. Residues, especially corn stover, are expected
to be the first feedstocks for cellulosic biofuels to be utilized. Excessive residue removal can
have important non- greenhouse gas environmental effects, such as erosion, and so should be
closely limited to a sustainable level. At any level, residue removal is likely to marginally
increase crop fertilizer needs and decrease soil organic carbon loads, resulting in some
greenhouse gas costs. Residues collected at processing sites, such as vegetable processing and
milling wastes, do not increase agricultural GHG emissions.
Municipal Solid Waste
Municipal solid waste ( MSW) destined for the landfill contains substantial ligno- cellulosic
material that can be converted to biofuels. The organic fraction of MSW capable of serving as a
biofuel feedstock ( which does not include plastics or other energy- rich materials) constitutes
55% of all MSW destined for the landfill in California ( Cascadia, 2004).
MSW, like industry residues, is already collected and concentrated, and so has a nearly- zero
production “ cost” and a low transportation cost.
6 Though, if perennial biomass feedstocks replace native ecosystems, there generally will be a
net carbon emission, not a net sequestration.
30
Oils
Oilseed crops, including soybeans, canola and mustard seeds, and sunflower seeds, are grown
throughout the United States. Palm oil is grown in tropical Southeast Asia and has been linked
to deforestation and the draining of peat bogs, both activities that result in large net GHG
releases. Some varieties of algae are known to produce large amounts of fatty acids and have
been proposed as biofuel feedstocks.
Biofuel Conversion Processes
The primary biofuels produced at a commercial scale in the U. S. today are fermented corn
ethanol and transesterified soybean biodiesel. Globally, fermented sugar cane ethanol is
produced in large quantities as well. Ethanol can be produced via simple fermentation from
other starch and sugar based feedstocks or from a wider range of cellulose based feedstocks
through an enzymatic process known as saccharification that releases the starches in cellulose.
Ethanol can also be produced from biomass- derived synthesis gas. Biodiesel can be produced
from a range of oil- based feedstocks via the currently predominant trans- esterification process.
Additional fuels in pilot- or small- scale applications include other alcohols ( e. g. biobutanol and
methanol), other diesel blendstocks ( e. g. Fischer- Tropsch fuels, renewable diesel, and dimethyl
ether), and gaseous biofuels ( e. g. hydrogen and methane).
Fermentation
Alcohols are generally produced through fermentation. While fermentation of the simple sugars
pressed from sugarcane, sweet sorghum, and sugar beets is simple, starch and cellulosic
materials require increasingly complicated ( and expensive) hydrolysis and saccharification
processes before sugars are available to fermentation. Fermentation with different yeasts can
produce either ethanol or butanol fuels.
Cellulosic material is often bound up with lignin in complex ways that must be broken before the
cellulose is available for saccharification and fermentation. Thus ligno- cellulosic material must
be pre- treated through one of several processes before enzymatic breakdown of cellulose can
occur. Candidate pre- treatment processes include dilute- acid pretreatment and ammonia- fiber-explosion.
The cost of pre- treatment is a major barrier to cellulosic alcohol production via the
saccharification- fermentation pathway.
Transesterification
The reaction of biomass oils with alcohol in the presence of a catalyst produces esters and
glycerin. The esters have similar properties to diesel, and glycerin is valuable coproduct.
Gasification
The partial combustion of biomass in an oxygen- limited environment can produce a CO- and
H2- rich gas called synthesis gas that can in turn be used in several processes to produce heat,
electricity, and liquid fuels. Synthesis gas can be reformed using the Fischer- Tropsch process to
hydrocarbons, primarily middle distillates for diesel production but also some gasoline
components. Synthesis gas can also be fermented to ethanol, or refined to a pure hydrogen fuel
product.
31
Flash Pyrolysis
Pyrolysis is the first stage of the gasification process that, when optimized by a short residence
time and zero- oxygen environment, produces in addition to combustible synthesis gas a heavy
liquid hydrocarbon called bio- oil. Bio- oil can be refined to gasoline- and diesel- like
hydrocarbons.
Hydrothermal liquefaction
Hydrothermal liquefaction uses high temperatures and pressure to combine water and biomass
and convert both to an oily liquid that can then be separated to hydrocarbons and organic- rich
water. The hydrocarbon components can then be added to standard petroleum feedstocks in
refinery operations.
32
Bibliography
1. 2005. Agricultural Statistics 2005. Rep. ISBN O- 16- 036158- 3, USDA, National
Agricultural Statistics Service, Washington, DC
2. 2006. Municipal Solid Waste in the United States, 2005 Facts and Figures. Rep.
EPA530- R- 06- 11, United States Environmental Protection Agency
3. 2007. Theoretical Ethanol Yield Calculator. US Department of Energy
4. Erickson GE, Klopfenstein TJ, Adams DC, Rasby RJ. 2005. Utilization of Corn Co-
Products in the Beef Industry, Nebraska Corn Board and University of Nebraska,
Lincoln, Lincoln, NE
5. Espey M. 1998. Gasoline demand revisited: an international meta- analysis of elasticities.
Energy Economics 20: 273- 95
6. Farrell AE, Plevin RJ, Turner BT, Jones AD, O'Hare M, Kammen DM. 2006. Biofuels
Can Contribute to Energy and Environmental Goals. Science 311
7. Greene N. 2004. Growing Energy: How Biofuels Can Help End America’s Oil
Dependence, National Resources Defense Council
8. HAYNES RW. 2003. An Analysis of the Timber Situation in the United States, 1952-
2050. Portland, OR: U. S. Department of Agriculture, Forest Service. 254 pp.
9. Hoffman L, Baker A, Foreman L, Young CE. 2007. Feed Grains Backgrounder, USDA
Economic Research Service
10. Koplow D. 2006. Biofuels - At What Cost? Governement support for ethanol and
biodiesel in the United States, The Global Subsidies Initiative ( GSI), Cambridge, MA
11. Liebig MA, Johnson HA, Hanson JD, Frank AB. 2005. Soil carbon under switchgrass
stands and cultivated cropland. Biomass and Bioenergy 28: 347- 54
12. Lubowski RN, Vesterby M, Bucholtz S, Baez A, Roberts MJ. Major Uses of Land in the
United States, 2002. Rep. EIB- 14, USDA Economic Research Service
13. Mann L, Tolbert V, Cushman J. 2002. Potential environmental effects of corn ( Zea mays
L.) stover removal with emphasis on soil organic matter and erosion. Agriculture,
Ecosystems & Environment 89: 149
14. McLaughlin SB, De La Torre Ugarte DG, Garten CT, Jr., Lynd LR, Sanderson M, et al.
2002. High- Value Renewable Energy from Prarie Grasses. Environ. Sci. Technol. 36:
2212- 129
15. McLaughlin SB, Kszos LA. 2005. Development of switchgrass ( Panicum virgatum) as a
bioenergy feedstock in the United States. Biomass and Bioenergy 28: 515- 35
16. Nelson RG, Walsh M, Sheehan JJ, Graham R. 2004. Methodology for Estimating
Removable Quantities of Agricultural Residues for Bioenergy and Bioproduct Use.
Applied Biochemistry and Biotechnology 113: 13- 26
17. Patzek T. 2006. The Earth, Energy, and Agriculture. Presented at Climate Change and
the Future of the American West - Exploring the Legal and Policy Dimensions, Boulder,
CO
18. Perlack RD, Wright LL, Turhollow A, Graham RL, Stokes B, Erbach DC. 2005. Biomass
as feedstock for a bioenergy and bioproducts industry: the technical feasibility of a
billion- ton annual supply. Rep. DOE contract: DE- AC05- 00OR22725, Oak Ridge
National Laboratory, Oak Ridge, TN
19. Preston RL. 2004. Typical Composition of Feeds for Cattle and Sheep, 2004. Beef: 20- 7
33
20. Puller SL, Greening LA. 1999. Household adjustment to gasoline price change: an
analysis using 9 years of survey data. Energy Economics 21: 37- 52
21. Simmons P, Goldstein N, Kaufman SM, Themelis NJ, James Thompson J. 2006. The
State of Garbage in America. BioCycle: 26- 43
22. Wang MQ. 1999. GREET 1.5 — Transportation Fuel- Cycle Model, Volume 1:
Methodology, Development, Use, and Results. Rep. ANL/ ESD- 39, Vol. 1, Center for
Transportation Research, Energy Systems Division, Argonne National Laboratory
23. Wilhelm WW, Johnson JMF, Hatfield JL, Voorhees WB, Linden DR. 2004. Crop and
Soil Productivity Response to Corn Residue Removal: A Literature Review. Agron J 96:
1- 17