1. No category

Applied Energy 127 (2014) 44–54
Contents lists available at ScienceDirect
Applied Energy
journal homepage: www.elsevier.com/locate/apenergy
Cost to produce and deliver cellulosic feedstock to a biorefinery:
Switchgrass and forage sorghum
Andrew P. Griffith a,1, Mohua Haque b,2, Francis M. Epplin c,⇑
a
Department of Agricultural and Resource Economics, University of Tennessee, 314B Morgan Hall, 2621 Morgan Circle, Knoxville, TN 37996, USA
Samuel Roberts Noble Foundation, 2510 Sam Noble Parkway, Ardmore, OK 73401, USA
c
Department of Agricultural Economics, Oklahoma State University, Stillwater, OK 74078-6026, USA
b
h i g h l i g h t s
We model field-to-biorefinery dedicated energy crop production and delivery cost.
We determine cost to produce and deliver switchgrass and forage sorghum biomass.
Estimated costs of delivering a flow of switchgrass is less than for forage sorghum.
The cost difference is primarily due to differences in harvest costs.
Harvest cost are influenced by the length of the harvest window.
a r t i c l e
i n f o
Article history:
Received 23 May 2012
Received in revised form 20 August 2013
Accepted 27 March 2014
Keywords:
Costs
Just-in-time
Logistics
Forage sorghum
Integer programming
Switchgrass
a b s t r a c t
Switchgrass and forage sorghum have both been proposed as potential candidates for high yielding, dedicated energy crops. This research was conducted to determine and compare the costs to produce and
deliver switchgrass and forage sorghum biomass under the assumptions that the biomass would be baled
and transported by truck and that the biorefinery would use either switchgrass or forage sorghum but not
both. A multi-region, multi-period, monthly time-step, mixed integer mathematical programming model
is used to determine the costs to deliver a flow of biomass to a biorefinery. The model is designed to
determine the optimal location of a biorefinery that requires 3630 Mg of biomass per day, the area
and quantity of feedstock harvested in each county by land category, the number of harvest machines
required, and the costs to produce, harvest, store, and transport a flow of biomass to a biorefinery. The
estimated costs of land rent, establishment, maintenance, fertilizer, harvest, storage, and transportation
is $60 Mg1 for switchgrass and $74 Mg1 for forage sorghum. The cost difference between the two crops
is primarily due to harvest costs, which are estimated to be $13 Mg1 greater for forage sorghum. Forage
sorghum has a narrower harvest window, requires more time for field drying prior to safe baling and, as a
consequence, requires significantly more harvest machines. Based on the assumptions used in this study
for Oklahoma conditions, a switchgrass system with a nine-month harvest window can deliver baled biomass at a lower cost than a forage sorghum system with a five-month harvest window. However, the
value of a Mg of switchgrass relative to a Mg of forage sorghum remains to be determined.
Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction
The U.S. Energy Independence and Security Act of 2007 (EISA)
mandates that U.S. retailers sell 136 billion L yr1 of biofuels by
the year 2022 (if they are produced), with 79 billion L yr1
⇑ Corresponding author. Tel.: +1 405 744 6156; fax: +1 405 744 8210.
E-mail addresses: [email protected] (A.P. Griffith), [email protected] (M.
Haque), [email protected] (F.M. Epplin).
1
Tel.: +1 865 974 7480.
2
Tel.: +1 580 223 5810.
http://dx.doi.org/10.1016/j.apenergy.2014.03.068
0306-2619/Ó 2014 Elsevier Ltd. All rights reserved.
expected to be forthcoming from lignocellulosic feedstocks such
as urban waste, forest biomass, and biomass from dedicated energy
crops [1]. The U.S. Billion-Ton Update reported that 16–24 million ha of cropland and pasture could be converted to produce
energy crops [2]. Switchgrass (Panicum virgatum L.) was evaluated
as the model perennial grass energy crop species, and forage sorghum (Sorghum bicolor L. Moench) was considered as the model
annual energy crop. Evaluating the logistics required to provide a
flow of biomass produced by the energy crops throughout the year
to a biorefinery was beyond the scope of the model used for the
Billion-Ton Update [2].
45
A.P. Griffith et al. / Applied Energy 127 (2014) 44–54
Switchgrass is considered a potential dedicated perennial
energy crop in the U.S. Southern Great Plains because, in that
region, it produces greater biomass yields than other warm season
grasses such as Kleingrass (Panicum coloratum L.), Johnsongrass
(Sorghum halepense L. Pers), and Bermudagrass (Cynodon dactylon
L. Pers) [3]. Miscanthus (Miscanthus x giganticus) has been found
to produce greater biomass yields than upland varieties of switchgrass in Illinois [4]. However, Aravindhakshan et al. [5] found that
lowland switchgrass varieties produced 28% more annual biomass
than miscanthus in a study conducted in the Southern Plains.
Forage sorghum has been proposed as an annual energy crop
because it has broad genetic diversity that provides the opportunity to develop varieties adapted to diverse climates [6]. It has
several desirable characteristics such as high yield potential,
water-use efficiency, drought tolerance, established production
systems, and the potential for genetic improvement using traditional and genomic approaches [7]. Hallam et al. [8] found that
in Iowa, forage sorghum produced more biomass than alternatives
that included reed canarygrass (Phalaris arundinacea L.), switchgrass, big bluestem (Andropogon gerardii Vitman), alfalfa (Medicago
sativa L.), and corn (Zea mays L.).
A number of studies have evaluated the farm gate costs of producing switchgrass [8–10] and forage sorghum [11–13]. For the
most part, these studies have ignored the logistics of transporting
a continuous flow of feedstock throughout the year from fields or
storage to a biorefinery. The costs incurred to move biomass from
the farm gate to provide a daily flow of feedstock to the biorefinery,
may differ substantially across feedstocks. In Oklahoma, switchgrass harvest may begin in July and extend though March. During
the nine-month harvest window, switchgrass biomass could be
harvested and delivered ‘‘just-in-time’’ (JIT). Preliminary estimates
are that the harvest window for forage sorghum in Oklahoma could
only extend for five months, from October through February. More
storage would be required for forage sorghum. A JIT system has
several advantages in that it could substantially reduce the investment required in harvest machines and reduce the amount of
space required for storage. However, a JIT system also has several
potential disadvantages. Expected switchgrass and forage sorghum
harvestable yields and expected fertilizer requirements differ
depending on the month of harvest [14] (Table 1). Harvesting
switchgrass prior to full maturity is expected to result in lower harvestable yields and greater fertilizer requirements. A comprehensive evaluation of a JIT system is needed to compare the
tradeoffs among yield, fertilizer, harvest machinery, storage, and
other factors.
This research attempts to compare the economic competitiveness of the proposed annual energy crop, forage sorghum, relative
to the proposed perennial crop, switchgrass. Separate models are
built for switchgrass and forage sorghum. The most economical
commercial scale system for converting lignocellulosic biomass
to economically competitive biobased products has not been
determined. Some studies model an enzymatic hydrolysis process.
Others model thermochemical processes such as gasification or
pyrolysis. It remains to be determined which of these several competing technologies will ultimately prevail, and if a biorefinery can
use multiple feedstocks. Since the harvest windows for forage sorghum and switchgrass overlap, potential economies from using
both feedstocks are not evident. Differences in the value of a delivered dry unit of switchgrass biomass relative to a dry unit of forage
sorghum also remain to be determined. However, the profitability
of a biorefinery will depend in part on the cost of delivered feedstock. The objective of this research is to determine and compare
the costs to deliver a year round flow of baled biomass to a biorefinery for both a system that uses forage sorghum exclusively and a
system that uses switchgrass exclusively. This type of information
will be essential to determine the potential economic viability of
biorefineries that plan to use either forage sorghum or switchgrass
biomass feedstock.
2. Methods
Since an infrastructure for producing and marketing biomass
feedstock does not exist, and since biomass feedstock has few
alternative uses, the risk would be very high for a biorefinery to
rely on spot markets for feedstock. To overcome some of these
risks, a biorefinery could develop a vertically integrated system
similar to that used by several U.S. wood products companies that,
through either ownership or leases, have rights to thousands of
hectares of timber land [15]. These companies manage timber production, harvest, transportation, processing, and sales of produced
products. A biorefinery that requires year round delivery of biomass could also be efficiently organized with a vertically integrated business plan [16,17].
Similar to integrated timber companies, production, harvest,
storage, and delivery of feedstock could be centrally managed
and coordinated. Land could be leased and seeded to energy crops
in a manner similar to what occurred when millions of U.S. ha were
enrolled in the Conservation Reserve Program (CRP) and seeded to
perennial grasses. The difference being that the biorefinery, rather
than the government, would be the lessee and would be responsible for paying the leasing costs. This system has the potential to
quickly identify and reduce bottlenecks and achieve cost efficiencies by managing quality throughout the field-to-products chain.
The optimal size of a cellulosic biorefinery is not known, but
economies of scale suggest the industry will ‘‘be characterized by
regionally dominant, large capacity biorefineries’’ [18]. Kazi et al.
[19] and Wright et al. [20] budgeted for 2000 dry Mg per day.
Wright and Brown [21] find that for some conversion technologies
optimally sized lignocellulosic biorefineries would require more
than 12,000 Mg per day. Regardless of the average feedstock yield,
a substantial quantity of land would be necessary to fulfill the
Table 1
Switchgrass and forage sorghum yield proportion and fertilizer requirements by harvest month.
January
a
July
August
September
October
November
December
Switchgrass
Forage sorghum
Proportion of potential yield by harvest montha
0.80
0.75
0.70
0.80
0.75
February
March
April
May
June
0.79
0.86
1.00
1.00
1.00
0.90
0.90
0.85
0.85
Switchgrass
Forage sorghum
Level of nitrogen (kg N ha1) by harvest month
71
71
71
101
101
90
83
77
71
101
71
101
71
101
Switchgrass
Forage sorghum
Level of phosphorus (kg P2O5 ha1) by harvest month
0
0
0
50
50
11
11
11
0
50
0
50
0
50
Switchgrass harvest is not permitted in April, May, and June. Forage sorghum harvest is not permitted from March through September.
46
A.P. Griffith et al. / Applied Energy 127 (2014) 44–54
needs of an efficiently sized biorefinery that requires feedstock
from dedicated energy crops and operates year round. A vertically
integrated system designed to maximize profit would include the
most economically efficient plan for harvesting and delivering a
flow of feedstock to an optimally located conversion facility year
round.
For a given annual biomass requirement, the number of harvest
machines required would depend on the length of the harvest window, the weather, and the number of harvest days. Extending the
harvest window could reduce the investment required in harvest
machines necessary to supply the biorefinery and also reduce the
amount of storage space needed. Extending the harvest window
would require additional land for growing feedstock due to the
lower average yield that is expected with an extended harvest window (Table 1). Harvesting switchgrass before dormancy could also
require more nitrogen because studies have determined that if harvest is delayed until after senescence, some proportion of the
nutrients translocate from the foliage to the crown and rhizomes
[22,23]. However, annuals such as forage sorghum do not translocate nutrients to the root system to the same degree as switchgrass
[24].
3. Model
A mixed integer mathematical programming model is constructed to determine the costs to lease land, plant, manage, harvest, and deliver a flow of biomass throughout the year. Land
acquisition, length of harvest window, suitable harvest days, and
the number of harvest machines required pose critical cost and
data issues.
3.1. Land use
The case study model includes all 77 Oklahoma counties as
individual production regions plus 11 potential biorefinery locations. Cropland and improved pasture land area for each county
are based on data from the Census of Agriculture [25]. The model
is constructed so that switchgrass can be produced on both cropland and improved pasture land, whereas forage sorghum production is limited to cropland. Forage sorghum, which must be
replanted each year, is restricted to cropland due to the increased
risk of soil erosion when grown on marginal lands. There are also
soil erosion concerns with switchgrass production in the establishment year if conventional tillage is used to prepare the seedbed
[26]. Switchgrass is a bunch grass, and until it becomes well established, does not provide substantial ground cover between plants.
However, in the Southern Plains as switchgrass matures and the
root system develops, tillers develop and provide substantial
ground cover. This reduces the risk of soil erosion relative to
annual crops such as forage sorghum if switchgrass is harvested
by mowing at a height of 15 cm or greater [24].
The model limits biomass production in a production region by
restricting area usage to no more than 10% of a county’s cropland
and no more than 10% of a county’s improved pasture land. The
assumption is made that cropland could be acquired for a longterm lease rate above average CRP rental rates [27]. The lease rate
for cropland for each county is calculated by adding a fixed amount
of $49 ha1 to the average CRP rental rate for that county as determined by Fewell et al. [28]. Long term lease rates for improved pasture land are derived by adding $76 ha1 to the 2010 average
county pasture rental rate [29]. The modeled rental rates are
designed to cover the opportunity costs of alternative production
options and to account for increased land-lease rates that may
occur in response to an entrant in the market for 10% of the
county’s land, and to compensate for the lost option value from
engaging in long term leases [30].
Switchgrass and forage sorghum biomass yield estimates for
each production region are obtained from estimates produced by
the Oak Ridge National Laboratory [2,31]. Switchgrass harvested
in July results in a lower expected yield and a greater expected
nitrogen requirement than switchgrass harvested in October
(Table 1). Switchgrass yield response to fertilizer for alternative
harvest months is based on data from a multiyear field trial [32].
3.2. Length of harvest window
Switchgrass is modeled as having a harvest window starting in
July and ending in March with no harvest expected in April, May, or
June due to potential damage to future year’s plant growth. Forage
sorghum is modeled with harvest to start in October and end in
February. Forage sorghum harvest is delayed until October because
it is a later maturing species. Both crops are assumed to be mowed,
dried in the field to no more than 15–20% moisture, and baled into
large rectangular solid bales. Forage sorghum harvest is modeled to
end in February because it has a higher incidence of lodging which
makes it more difficult to harvest [33].
3.3. Suitable harvest days
The moisture content of the biomass should be no more than
15–20% when baled because the higher the moisture content when
baled, the higher the incidence of mold, premature fermentation,
and potential spontaneous combustion. Time after mowing is
required to enable the material to dry prior to baling. Drying is
enhanced by the relatively low humidity of the Southern Plains.
However, forage sorghum has a propensity to retain moisture
and more dry down time is expected to be required for forage sorghum than for switchgrass [34,35].
Hwang et al. [36] uses historical weather data to model time
required for switchgrass biomass to dry between mowing and baling and to determine distributions of suitable switchgrass harvest
days by month for Oklahoma counties. In most months, the number of suitable mowing days exceeds the number of suitable
switchgrass biomass baling days. Similar harvest day by month
distributions are not available for forage sorghum for the region.
Given the lack of more precise data, the forage sorghum model is
solved under two different assumptions regarding dry down time
between mowing and baling and harvest days per month. For the
base model, forage sorghum is assumed to require twice as much
dry down time as switchgrass. For the alternative, dry down time
is assumed to be the same for forage sorghum as for switchgrass.
Findings from these models can provide information regarding
the relative cost of providing biomass from the two species in a
baled form.
3.4. Harvest machines
Biomass harvest costs are based on the integrated harvest unit
concept originally developed by Thorsell et al. [37]. Machinery
requirements for harvest include machines for mowing, raking,
baling, and stacking. A self-propelled windrower (140 kW) with a
4.9 m rotary header and a driver is modeled as the mowing unit.
The harvest unit modeled for raking, baling, field transport, and
stacking consists of three wheel rakes, three 40 kW tractors, three
balers, three 147 kW tractors, a field transporter, and seven workers. The modeled wheel rake consists of two 3 m rakes pulled in
tandem. The modeled baler constructs 1.2 m 1.2 m 2.4 m solid
rectangular bales. The self-propelled field transporter collects as
many as eight bales, transports them, and stacks them at a location
47
A.P. Griffith et al. / Applied Energy 127 (2014) 44–54
near an all-weather road. The mowing unit and harvest unit are
included in the model as integer variables.
3.5. Transportation
The most efficient and cost-effective method to harvest, store,
and transport massive quantities of biomass feedstock to a central
location, has been the subject of research studies for a number of
years [38]. In general the studies have found that the most economically efficient harvest, storage, and transportation system
depends on assumptions regarding field size, climate, length of
harvest window, transportation distance, and form of the delivered
material [39–43]. For example, Worley and Cundiff [44] report that
field chopping is not cost competitive with baling, even in the very
humid climate of central Florida. Sultana and Kumar [45] model
transportation cost for delivering material to a pelleting facility.
Judd et al. [46] found that large cylindrical bales would be a more
economical form than rectangular solid bales given the relatively
small fields and more humid Southeast region of the USA. Kumar
and Sokhansanj [47] compared several potential harvest and transport systems. However, they assume only a four month harvest
window and that the cost of the delivered material include a cost
for grinding. They find that an experimental loafing system would
be more economical than baling but caution that unlike baling,
their loafing system has not been validated at a commercial scale.
For climate and field conditions of the Southern Plains, Thorsell
et al. [37] concluded that if the search for system type was limited
to existing forage harvest system technologies, the rectangular
solid bale system would be most economical. Bales could be harvested and stored in or close to the harvest field near an allweather road. Bales could then be trucked as necessary to the
biorefinery.
Wang [48] estimated the cost of transporting baled biomass by
truck from fields where produced to a biorefinery. The model
assumes a semi-tractor trailer is used to transport biomass in the
form of solid rectangular bales (1.2 m 1.2 m 2.4 m) from the
field’s edge to the processing facility. The trailer is designed to
carry 24 bales weighing approximately 0.9 Mg each. At 15–20%
moisture a load would contain approximately 18 Mg dry biomass.
Wang’s [48] transportation cost estimate can be described as a
function of the price of diesel fuel and the travel distance between
the field of production and the biorefinery. The equation is:
TC = 0.8799 + [0.1648Pd + 0.0677]km, where TC is the round trip
cost ($ Mg1) to transport one dry Mg of baled biomass from a distance of km kilometers; Pd is the price ($ L1) of diesel fuel and km
is the transportation distance from the field to the biorefinery. The
equation accounts for the total round trip cost.
machines, and the costs to produce, harvest, store, and transport
a continuous flow of biomass to a biorefinery. The objective function is constructed to maximize the net present value (NPV) of
the system:
maxNPV
Q jm ;Ailm ;XT ijkm ;XSIPikm ;X ilm ;XPjkm ;XSIikm ;XSNikm ;XSJjkm ;HUM;HUB
(
J
M
X
X
¼
m¼1
I X
K
L
I X
K
L
X
X
X
X
dk Ailm gik Ailm
i
j¼1
k
i
l¼1
k
l¼1
I X
K
X
L
I X
L
I X
L
X
X
X
mik Alim alm Ailm clm Ailm
i
qQ jm k
l¼1
J X
I X
K
X
i¼1 l¼1
sij XT ijkm i
j
k
i¼1 l¼1
!
I X
K
X
Ck XSIP ikm i¼1 k¼1
xHUM -HUBg PVAF J X
F
X
OMC f bj
j¼1 f ¼1
J X
F
X
AFC f bj
ð1Þ
j¼1 f ¼1
PJ
where
qQ jm is the return from biobased products,
PI PK PLj¼1
d
A
i
k k
l¼1 ilm is the cost of producing bioenergy crops excluding
PI PK PL
fertilizer and harvest,
i
k gik
l¼1 Ailm is establishment cost,
PI PK PL
PI PL
m
A
is
land
rent,
i
k ik
l¼1 ilm
i¼1
l¼1 alm Ailm is nitrogen cost,
PI PJ PK
PI PL
i¼1
l¼1 clm Ailm is phosphorus cost,
i
k sij XT ijkm is transportaj
tion cost. sij = 0.8799 + [0.1648Pd + 0.0677]kmij, where Pd is the
price ($ L1) of diesel fuel and kmij is the distance from county i
PI PK
to biorefinery location j.
i¼1
k¼1 Ck XSIP ikm is storage cost,
PJ PF
OMC
b
is
operating
and
maintenance
cost, xHUM is harf j
f ¼1
j¼1
vest cost for mowing, -HUB is harvest cost for raking–baling–stacking operations, PVAF is the present value of annuity factor,
PJ PF
f ¼1 AFC f bj is biorefinery investment cost, and Tables 2 and 3
j¼1
include descriptions of set member elements, parameters, and
variables.
PVAF ¼
ð1 þ rÞT 1
ð2Þ
rð1 þ rÞT
alm ¼ Pn NIT lm
ð3Þ
clm ¼ PP PIT lm
ð4Þ
Eq. (1) is maximized subject to a set of constraints. Eq. (5) restricts
total planted switchgrass or forage sorghum in each county on cropland to not exceed a set proportion of the quantity of available cropland. In this study, BIPROP was set to 10% therefore limiting the
quantity of cropland bid from traditional crops to produce dedicated energy crops to 10% of total cropland in the county.
3.6. The model
M
X
Building on and extending the work of Tembo et al. [49] and
Mapemba et al. [50], multi-region, multi-period, monthly timestep, mixed integer mathematical programming models are formulated and used to determine the cost to deliver a steady flow of biomass to a biorefinery [51,52]. This research was conducted to
determine and compare the costs to produce and deliver switchgrass and forage sorghum biomass under the assumptions that
the biomass would be baled and transported by truck and that
the biorefinery would use either switchgrass or forage sorghum
but not both.
The models are designed to determine the optimal location of a
biorefinery that requires a flow of 3630 Mg of biomass per day
from among 11 locations, Blaine, Canadian, Carter, Garfield, Grady,
Kay, Okmulgee, Payne, Pontotoc, Washington, and Woods counties.
The models also determine the area and quantity of feedstock harvested in each county by land category, the number of harvest
m¼1
Ailm BIPROP POTACREil 6 0;
8i ; l ¼ cropland
ð5Þ
Similar to cropland, Eq. (6) restricts total planted switchgrass on
improved pasture land in each county by setting BIPROP1 to 10%.
M
X
Ailm BIPROP1 POTACREil 6 0;
8i ; l ¼ improved pasture land
m¼1
ð6Þ
Eq. (7) is a yield balance equation used to calculate the amount of
biomass produced on harvested lands.
L
L
X
X
X ilm Ailm BYLDil YADkm ¼ 0;
l¼1
8i;k;m
ð7Þ
l¼1
Eq. (8) limits harvest months. The model sets YADkm equal to zero in
the months of April, May, and June for switchgrass and the months
48
A.P. Griffith et al. / Applied Energy 127 (2014) 44–54
Table 2
Description of sets and variables used in the models.
Symbol
Description
Sets
M
J
I
F
K
Months: m = {January, February, March, April, May, June, July, August, September, October, November, December}
Prospective biorefinery locations: j = {Blaine, Canadian, Carter, Garfield, Grady, Kay, Okmulgee, Payne, Pontotoc, Washington, Woods}
Biomass source counties: i = {77 Oklahoma counties}
Facilities: f = {processing, storage}
Switchgrass or forage sorghum production system: k = {established on cropland, established on improved pasture land}
Land class: l = {cropland, improved pasture land}
Variables
NPV
Qjm
Ailm
XSIikm
XSIPikm
XTijkm
HUM
HUB
Xilm
XPjkm
XSIikm
XSJjkm
XSINim
XHUMim
XHUBim
bj
Net present value of the system ($)
Quantity of ethanol produced in month m by a biorefinery at location j (L)
Land harvested in month m from land class l in county i (hectares)
Biomass stored in field in month m from system k in county i (Mg)
Biomass placed into storage in month m from system k in county i (Mg)
Biomass transported from county i in month m from system k to a biorefinery at location j (Mg)
Integer variable representing the total number of mowing harvest units
Integer variable representing the total number of raking-baling-stacking harvest units
Biomass harvested in month m from land class l in county i (Mg)
Biomass processed in month m from system k at location j (Mg)
Biomass stored as source i from system k in month m (Mg)
Biomass stored in month m from system k at location j (Mg)
Biomass removed from field storage in month m and county i (Mg)
Proportion of a harvest unit for mowing used in month m in county i
Proportion of a harvest unit for raking-baling-stacking used in month m in county i
Binary variable for biorefinery location j (1 if built, 0 otherwise)
of March through September for forage sorghum, resulting in no
harvest during the respective months.
I
X
XT ijkm þ hJ k XSJ jkm1 XSJ jkm XP jkm ¼ 0;
8j;k;m
ð14Þ
i¼1
L
X
Ailm ¼ 0 if YADkm ¼ 0;
8i;k;m
ð8Þ
l¼1
The sum of biomass transported to the plant location from each
county, in addition to stored biomass, is set to be equivalent to
the sum of current production and the usable portion of stored biomass at the source county for each month by Eq. (9).
J
L
X
X
X ilm þ hIk XSIikm1 XT ijkm XSIikm ¼ 0;
8i;k;m
ð9Þ
j¼1
l¼1
Eq. (10) equates total biomass quantity transported to the biorefinery plus the storage loss to quantity harvested.
L X
M
X
J X
M
M
X
X
X ilm XT ijkm ð1 hIk Þ XSIikm ¼ 0;
l¼1 m¼1
j¼1 m¼1
8i;k
ð10Þ
The total quantity of biomass delivered from each production
region to the biorefinery is equated to the total quantity of processed biomass plus storage losses as shown in Eq. (15).
I X
M
X
i¼1 m¼1
M
X
m¼1
K
X
XSJjkm BINVbj P 0;
Q jm K
X
kk XP jkm 6 0;
8m
ð11Þ
Eq. (12) limits monthly biorefinery processing capacity for each
location.
Q jm CAPPbj 6 0;
8j;m
ð12Þ
8j;m
X
XSJ jkm CAPbj 6 0;
8j;m
ð13Þ
k¼1
Eq. (14) restricts the quantity of biomass transported to the biorefinery in a specific month minus the quantity processed at the biorefinery during that month to be equal to the change in biomass
storage inventory during the month at the biorefinery.
ð17Þ
The number of endogenously determined mowing units used in any
month is restricted to not exceed the available number of mowing
units (Eq. (18)).
8m
ð18Þ
The number of raking–baling–stacking harvest units used in any
month is restricted by Eq. (19) to not exceed the total number of
raking–baling–stacking harvest units endogenously determined by
the model.
I
X
XHUBim HUB 6 0;
Eq. (13) limits monthly storage capacity at the biorefinery.
ð16Þ
k¼1
i¼1
¼ 0;
ð15Þ
Eq. (17) restricts ethanol production in each month to not exceed
the capacity of the biorefinery.
i¼1 l¼1
i¼1 k¼1
8j;k
m¼1
8j;m
I
X
XHUMim HUM 6 0;
i¼1 k¼1
M
X
XP jkm ¼ 0;
k¼1
J X
I X
L
I X
K
I X
K
I X
K
X
X
X
X
X ilm XT ijkm þ
XSINikm XSIPikm
i¼1 j¼1 k¼1
XSJ jkm A minimum biomass inventory at the biorefinery is imposed by Eq.
(16).
m¼1
The total quantity of harvested biomass plus the quantity of biomass removed from field storage each month is set equal to the
amount of biomass transported from each county to the biorefinery
plus the amount of biomass placed in storage at the biorefinery (Eq.
(11))
XT ijkm ð1 hJ k Þ
8m
ð19Þ
i¼1
Eqs. (20)–(23) ensure that each month’s harvested biomass is less
than the harvesting capacity of the total number of mowing harvest
units and raking–baling–stacking harvest units.
CAPHUMim ¼ FWDim DCAMHU m
8i;m
ð20Þ
A.P. Griffith et al. / Applied Energy 127 (2014) 44–54
49
Table 3
Descriptions of parameters used in the models.
Parameter
Description
q
Price of ethanol ($ L1)
Price of nitrogen ($ kg1)
Price of P2O5 ($ kg1)
Cost of producing switchgrass or forage sorghum with system k excluding cost of fertilizer and harvest ($ ha1)
Establishment cost for county i by production system k ($ ha1)
Land rent for county i by production system k ($ ha1)
Cost of applied nitrogen to land class l harvested in month m ($ ha1)
Cost of applied P2O5 to land class l harvested in month m ($ ha1)
Round-trip cost of transporting biomass from county i to biorefinery located j ($ Mg1)
Cost of storing biomass in the field with production system k ($ Mg1)
Annual cost of a mowing unit ($ per unit)
Annual cost of a raking-baling-stacking unit ($ per unit)
Usable proportion of biomass from production system k stored in field (1 – storage loss %)
Usable proportion of biomass from production system k stored at biorefinery (1 – storage loss %)
Quantity of ethanol produced from a ton of biomass from production system k (L Mg1)
Proportion of cropland in each county available for producing biomass
Proportion of improve pasture land in each county available for producing biomass
Nitrogen applied to land class l when harvested in month m (kg ha1)
P2O5 applied to land class l when harvested in month m (kg ha1)
Hectares of land class l in county i
Biomass yield adjustment factor for production system k harvested in month m
Biomass yield from production in county i on land class l (Mg ha1 yr1)
Biorefinery operating and maintenance cost for facility of size s type f ($ yr1)
Biorefinery investment cost for facility of size s type f made once in year 0 ($)
Present value of annuity factor
Plant life (years)
Discount rate (%)
Minimum biomass inventory for plant size s (Mg per month)
Processing capacity of biorefinery size s (L of ethanol per month)
Storage capacity of biorefinery size s (Mg of biomass)
Field work days suitable for mowing in county i in month m
Daily capacity of a mowing harvest unit in month m
Capacity of mowing harvest unit in month m
Field work days suitable for raking-baling-stacking in county i in month m
Daily capacity of a raking-baling-stacking harvest unit in month m
Capacity of raking-baling-stacking harvest unit in month m
Pn
PP
dk
gik
mik
alm
clm
sij
Ck
x
hIk
hJk
kk
BIPROP
BIPROP1
NITlm
PITlm
POTACREil
YADkm
BYLDil
OMCf
AFCf
PVAF
T
r
BINV
CAPP
CAP
FWDim
DCAMHUm
CAPHUMm
BWDim
DCABHUm
CAPHUBm
The monthly capacity of a mowing harvest unit is calculated by
multiplying the capacity of a mowing unit in month m by the number of mowing days available in month m.
4. Results
L
X
X ilm XHUMim CAPHUMim 6 0;
The models determine that the biorefinery would be optimally
located in Blaine County for both the switchgrass and forage sorghum production systems. A summary of estimated costs, optimal
number of windrowers and balers, quantity of land used, and feedstock harvested for supplying the biorefinery with 3630 Mg of
feedstock per day is provided in Table 4. The estimated costs of
land rent, establishment, maintenance, fertilizer, harvest, storage,
and transportation is $60 and $74 Mg1 for switchgrass and forage
sorghum respectively (Table 4; and Fig. 1). The cost difference
between switchgrass and forage sorghum is primarily a result of
differences in harvest costs, which are estimated to be $13 Mg1
greater for forage sorghum than for switchgrass.
The forage sorghum system requires 37 more windrowers for
mowing and 249 more balers, which increases machinery ownership costs. The optimal number of windrowers for mowing switchgrass is 34, while 71 are estimated to be needed for forage
sorghum. The optimal set of raking, baling, and stacking machines
for switchgrass includes 81 rakes, 81 balers, 162 tractors, and 27
field stackers. For forage sorghum, the optimal set includes 330
rakes, 330 balers, 660 tractors, and 110 field stackers. The requirement of additional harvest machines for forage sorghum follows
from the assumption of a narrower harvest window (five months
versus nine months for switchgrass) and the assumption that forage sorghum requires more dry down time resulting in only half
as many baling days per month. For four of the five forage sorghum
harvest months, at least twice as much forage sorghum is scheduled to be harvested as compared to switchgrass (Fig. 2).
8i;m
ð21Þ
l¼1
CAPHUBim ¼ BWDim DCABHU m
L
X
X ilm XHUBim CAPHUBim 6 0;
8i;m
ð22Þ
8i;m
ð23Þ
l¼1
Eq. (24) equates the raking–baling–stacking capacity in each production region and each month with the mowing capacity.
XHUMim CAPHUMim XHUBim CAPHUBim ¼ 0;
8i;m
ð24Þ
Eq. (25) lists non-negative decision variables. The number of mowing harvest units (HUM) and the number of raking–baling–stacking
harvest units (HUB) are set to be non-negative integer values.
Q jm ; Ailm ; XT ijkm ; X lim ; XP jkm ; XSIikm ; XSIP ikm ; X ilm ; XSINikm ;
XSJ jkm ; XHUM; and XHUB P 0
ð25Þ
Eq. (26) restricts the biorefinery location variable to be binary.
bj 2 f0; 1g
ð26Þ
The switchgrass model includes 10,900 equations and 75,000
activities.
4.1. Base scenario
50
A.P. Griffith et al. / Applied Energy 127 (2014) 44–54
Table 4
Estimated cost, number of windrowers and balers, area and quantity of biomass harvested for both switchgrass and forage sorghum for each scenario.
Category
Base scenario
Units
a
b
Mg1
Mg1
Mg1
Mg1
Mg1
Mg1
Mg1
Same dry-down
timea,b
Forage sorghum
Fuel price doubled
Land rent doubled
Switchgrass
Forage
sorghum
Switchgrass
Forage
sorghum
Yield
doubleda,b
Forage
sorghum
Switchgrass
Forage
sorghum
12.22
7.61
7.64
0.46
15.73
0.48
16.06
9.52
6.26
7.20
4.15
28.85
1.22
16.57
9.55
6.27
7.21
4.16
20.16
1.22
16.10
12.47
7.69
7.70
0.46
18.92
0.48
25.39
9.72
6.36
7.31
4.21
33.21
1.22
25.35
24.08
7.54
7.57
0.45
15.34
0.48
16.86
18.67
6.20
7.13
4.11
28.90
1.22
17.07
4.79
3.19
7.34
4.23
29.66
1.22
11.61
Land rent
Establishment and maintenance cost
Cost of nitrogen
Cost of phosphorus
Harvest cost
Field storage cost
Transportation cost
$
$
$
$
$
$
$
Total cost of delivered feedstock
Windrowers
Balers
Biomass harvested from cropland
Biomass harvested from improved
pasture land
Total biomass harvested
Cropland harvested
Improved pasture land harvested
$ Mg1
No.
No.
Mg
Mg
60.20
34
81
795,997
493,262
73.77
71
330
1,312,184
64.67
72
168
1,311,640
73.11
35
81
820,642
468,144
87.38
73
339
1,311,402
72.32
35
78
760,958
528,216
83.30
72
330
1,312,160
62.04
73
339
1,311,490
Mg
ha
ha
1,289,259
78,636
49,944
1,312,184
92,387
1,311,640
92,508
1,288,787
81,366
48,261
1,311,402
93,737
1,289,173
74,938
52,673
1,312,160
91,414
1,311,490
47,063
Total land harvested
ha
128,581
92,387
92,508
129,627
93,737
127,611
91,414
47,063
The number of suitable baling days per month for forage sorghum were set equal to the number of switchgrass baling days.
Yields in each county doubled for forage sorghum.
Forage sorghum requires fewer ha of land because the average
forage sorghum harvested yield is 14.2 Mg ha1 while switchgrass
averages about 10 Mg ha1 for the production regions selected by
the model. The forage sorghum system requires 92,387 ha of cropland whereas the switchgrass system requires 128,581 ha
(78,636 ha of cropland and 49,944 ha of improved pasture land).
Forage sorghum also requires less land because the narrower
harvest window results in less yield loss from leaving biomass
standing in the field (Table 1).
4.2. Sensitivity to number of baling days for forage sorghum
Safe baling requires that the mowed biomass contain no more
than 15–20% moisture. Hwang et al. [36] used historical weather
data and forage dry-down models to determine probability distributions of the number of days per month that switchgrass could be
safely baled in Oklahoma. Similar information is not available for
forage sorghum. For the base model, forage sorghum is assumed
to require twice as long to dry to safe baling moisture levels, and
thus baling days for forage sorghum are set at half the level as
for switchgrass. The number of harvest machines required and
the estimate of harvest costs depend critically on this constraint
on the number of forage sorghum harvest days per month. To test
the sensitivity of the findings, the number of harvest days for forage sorghum for each of the five harvest months is set equal to the
number of days modeled for switchgrass.
The cost to deliver forage sorghum decreases from $74 Mg1 to
$65 Mg1 when the constraint for the available number of baling
days during the October through February harvest window for forage sorghum is relaxed to be the same as that used for switchgrass
during these months. Under the base scenario, delivering forage
sorghum rather than switchgrass costs $14 Mg1 more. However,
when the number of available baling days constraint is relaxed
for forage sorghum, delivering forage sorghum costs $5 Mg1 more.
Relaxing the constraint on the number of baling days for forage
sorghum reduces harvest costs from $29 Mg1 (base scenario) to
$20 Mg1. Harvest costs are reduced by $9 Mg1 because 162 fewer
balers, 162 fewer mowers, 324 fewer tractors, and 54 fewer field
stackers are estimated to be required to harvest the crop during
the five month window.
4.3. Sensitivity to changes in fuel price
Table 4 also includes findings from doubling the price of diesel
fuel used to harvest and transport feedstock. The total cost to deliver switchgrass increases to $73 Mg1 while the total cost to deliver forage sorghum increases to $87 Mg1. An increase in fuel price
increases transportation costs, and the model optimally shifts
switchgrass production from improved pasture land that is further
from the biorefinery to cropland that is nearer the biorefinery
(Fig. 3). The optimal quantity of improved pasture land decreases
from 49,944 ha to 48,261 ha while cropland in biomass production
increases from 78,636 ha to 81,366 ha (Table 4).
An increase in fuel price would have relatively less effect on the
optimal location of forage sorghum production. Some shifts do
occur since cropland rental rates and expected yields vary across
counties. Feedstock production would optimally shift to counties
that are closer to the biorefinery with higher rental rates because
the greater transportation costs from shipping from more distant
county exceeds the added land costs from leasing land closer to
the biorefinery. An increase in the gap between delivered cost of
forage sorghum and switchgrass when the price of diesel fuel is
doubled, follows from the assumption that switchgrass can be
grown on both improved pasture land and cropland whereas forage sorghum is limited to cropland.
4.4. Sensitivity to changes in land rent
The optimal biorefinery location remains in Blaine County when
land rental rates are doubled. The estimated cost to deliver switchgrass increases by $12 Mg1 from $60 Mg1 (base scenario) to
$72 Mg1 while the cost to deliver forage sorghum increases by
$9 Mg1 from $74 Mg1 to $83 Mg1 (Table 4, and Fig. 1). Doubling
the land rental rate increases the estimated transportation costs of
switchgrass due to a shift in the type and location of land under production. Since switchgrass can be produced on both cropland and
improved pasture land, an increase in the land rental rate decreases
the amount of cropland and increases the amount of improved pasture land (Fig. 3). Fig. 3 illustrates that when the land rental rates
double, the production region extends, and fewer ha of cropland
and more ha of improved pasture land are optimally leased.
51
A.P. Griffith et al. / Applied Energy 127 (2014) 44–54
Fuel Price Doubled
Base Scenario
100
Transportation Cost
74
Field Storage Cost
60
Cost ($/Mg)
Cost ($/Mg)
80
60
Harvest Cost
40
Cost of Fertilizer
20
Establishment and
Maintenance Cost
0
Switchgrass
Forage Sorghum
100
90
80
70
60
50
40
30
20
10
0
Land Rent
Transportation Cost
80
Field Storage Cost
Cost ($/Mg)
Cost ($/Mg)
Establishment and
Maintenance Cost
Forage Sorghum
Forage Sorghum Yield Doubled
72
80
Cost of Fertilizer
100
Transportation Cost
83
Harvest Cost
Feedstock Source
Land Rent Doubled
100
Transportation Cost
Field Storage Cost
Switchgrass
Land Rent
Feedstock Source
87
73
60
Harvest Cost
40
Cost of Fertilizer
20
62
60
Field Storage Cost
60
Harvest Cost
40
Cost of Fertilizer
20
Establishment and
Maintenance Cost
0
Switchgrass
Forage Sorghum
Switchgrass
Land Rent
Feedstock Source
Establishment and
Maintenance Cost
0
Forage Sorghum
Feedstock Source
Land Rent
Same Number of Baling Days
100
Transportation Cost
Cost ($/Mg)
80
65
Field Storage Cost
60
60
Harvest Cost
40
Cost of Fertilizer
20
Establishment and
Maintenance Cost
0
Switchgrass
Forage Sorghum
Feedstock Source
Land Rent
Fig. 1. Estimated costs ($ Mg1) to provide a flow of feedstock throughout the year to a biorefinery for both switchgrass and forage sorghum.
$62 Mg1 (Table 4). The cost difference between forage sorghum
and switchgrass decreases from $14 Mg1 (base scenario) to
$2 Mg1 (Table 4). Harvest costs for forage sorghum are
$14 Mg1 greater than for switchgrass. However, when forage sorghum yields are doubled with switchgrass yields held constant,
forage sorghum has cost advantages because only 47,063 ha of
land are required. Land leasing and farming costs are decreased
accordingly. Transportation costs decline because average transportation distance declines with greater average yields.
Forage Sorghum
Switchgrass
Harvested Biomass (Mg)
350000
300000
250000
200000
150000
5. Discussion
100000
50000
Dec
Nov
Oct
Sep
Aug
Jul
Jun
May
Apr
Mar
Jan
Feb
0
Month of Harvest
Fig. 2. Biomass harversted by month for switchgrass and forage sorghum for the
base scenarios.
4.5. Sensitivity to changes in forage sorghum yield
When forage sorghum yields are doubled, the estimated cost to
deliver a flow of feedstock decreases from $74 (base scenario) to
Prior to investing hundreds of millions of dollars in a lignocellulosic biorefinery, due diligence would require a business plan that
encompasses the complete chain from feedstock acquisition to
sales of products produced. Spot markets for switchgrass and forage sorghum biomass do not currently exist. Rational land owners
would not enter into biomass feedstock production until a market
is available. A rational investor would not invest in a biorefinery
that did not have a reasonable plan for obtaining a flow of feedstock. One alternative would be for the biorefinery to engage in
long-term leases with land owners to acquire the rights to a sufficient quantity of land to produce feedstock. The models presented
in this paper follow from the assumption that feedstock production, harvest, and delivery is managed by the biorefinery to provide
a flow of biomass throughout the year.
52
A.P. Griffith et al. / Applied Energy 127 (2014) 44–54
Blaine County
Blaine County
Blaine County
Fig. 3. Cropland and improved pasture land usage for switchgrass under the base scenario, when fuel prices are doubled, and when land prices are doubled. One dot
represents 500 ha. Dots are randomly assigned within a county.
Based on the assumptions included in the model for the region
under study, the switchgrass system can deliver a year round flow
of biomass to a biorefinery at a lower cost than the forage sorghum
system. Differences in the value of a unit of biomass between the
species are not considered. Thus, these findings should not be
interpreted as a blanket endorsement of switchgrass over forage
sorghum. Differences exist in the components of switchgrass and
forage sorghum biomass, and thus the two feedstocks may not
be of equal value to the biorefinery. Research to determine the
optimal conversion system and optimal products will be required
to determine the most economically efficient field-to-biobased
products system. Though forage sorghum has a yield advantage
over switchgrass and the forage sorghum system requires less
nitrogen fertilizer per Mg, switchgrass has the advantage of a
longer harvest window, more harvest days in a harvest month,
and the ability to be produced on marginal land currently used
for improved pastures.
Results confirm land and fertilizer requirements are greater for
switchgrass than for forage sorghum, but following from the
assumption that the biomass be dried to safe baling moisture in
the field; the investment required for harvest machines is greater
for forage sorghum than for switchgrass. Harvest machinery
investment for forage sorghum is estimated to be more than three
times greater than the harvest machinery investment required for
switchgrass (base scenario). The difference in harvest machinery
investment accounts for much of the additional cost for delivering
A.P. Griffith et al. / Applied Energy 127 (2014) 44–54
a flow of forage sorghum biomass relative to switchgrass. This finding demonstrates the importance of the length of the harvest window, and the importance of the time required for mowed material
to dry to moisture levels required for safe baling. Harvest costs for
feedstocks produced in regions that enable longer harvest windows and an extended JIT system can be expected to be lower than
harvest costs of feedstocks produced in regions with relatively
short harvest windows.
This finding follows from the assumption that forage sorghum
would be harvested in a similar manner as switchgrass and with
the same type of equipment. Only baling was considered. Additional research would be required to determine if alternative harvest systems such as loafing or field chopping and ensiling would
be more economical than baling [44,47]. Optimal particle size
may differ across conversion system. Additional research would
be required to determine if baling, along with transportation and
storage of bales, and particle size reduction at the biorefinery, as
modeled here, is more economically efficient than field chopping
and transportation, storage, and processing of chopped material.
The optimal harvest system and the optimal field-to-biobased
products system may differ across species and climate.
The assumption that switchgrass could be produced on both
cropland and improved pasture land whereas forage sorghum
was restricted to cropland also benefits switchgrass. When rental
rates for cropland increase relative to those for improved pasture
land, production optimally shifts to more distant improved pasture
land. Likewise, if transportation costs increase, then production on
more distant improved pasture land will decrease, and production
on cropland closer to the biorefinery will increase. These shifts follow from relative changes in rental rates and transportation costs.
In the model, transportation costs are a function of biomass
transportation distance. These estimated travel distances are influenced by the assumption regarding biorefinery daily feedstock
requirements and the assumption that no more than 10% of a
county’s cropland and no more than 10% of a county’s improved
pasture land can be used. Average estimates of transportation costs
would be lower for a smaller biorefinery and they would be lower
if the 10% constraints were relaxed.
As shown in Table 4, both harvest and transportation costs are
sensitive to the price of diesel fuel. A diesel fuel price of $0.483 L1
was used for the base model and of $0.966 L1 for the fuel price
doubled model. In the U.S., excise taxes are charged on fuel purchases. These taxes are used to fund road and bridge maintenance
and construction. Fuel used for off road purposes (e.g. farming, logging, generating electricity) is exempt from these taxes. In some
states the off road excise tax exempt diesel fuel is dyed enabling
authorities to determine if a vehicle being used on the highway
has paid the tax. These excise taxes vary across state but average
about $0.15 L1 [53]. This is more than 30% of the budgeted base
fuel price. It remains to be determined if fuel used in biomass
production system would be exempted from these taxes. To be
consistent, in most states fuel used to produce and harvest would
be exempt. However, fuel used to transport biomass over the existing road system could be expected to be taxed. One limitation of
our model is that we assume the same price for fuel used for production and harvesting as for the transportation fuel. In any case, a
change in the net price of fuel used to power the machines used to
produce, harvest, and transport the biomass will change the cost to
deliver feedstock. But, based on the modeling assumptions, a
change in fuel price is not likely to change the cost advantage for
switchgrass relative to forage sorghum.
A facility capable of processing feedstocks with different physical, agronomic, and chemical characteristics could mitigate some
of the economic advantages that one species has over another.
Additional research will be required to determine if a field-tobiobased products system that uses a combination of feedstocks
53
would be more economical than a system that uses a single
feedstock. Since the harvest windows for forage sorghum and
switchgrass overlap, benefits from using a combination of the
two are not evident if both are required to be baled.
6. Conclusions
Switchgrass and forage sorghum have been identified as potential dedicated energy crops for cellulosic ethanol production. To
determine and compare the costs to deliver a year round flow of
biomass to a biorefinery, a multi-region, multi-period, monthly
time-step, mixed integer mathematical programming model is
developed for both switchgrass and forage sorghum. The model
determines the optimal biorefinery location, the area and quantity
of feedstock harvested in each county by land category, and the
number of mowing units and baling units necessary for harvest.
The model also provides an estimate of the costs to produce, harvest, store, and transport a continuous flow of biomass to a biorefinery. Based on the programming model, the estimated cost to
deliver a year round flow of switchgrass to a biorefinery is
$60 Mg1 while the estimated cost for forage sorghum is
$74 Mg1. The advantage switchgrass has over forage sorghum follows from the assumptions that (a) the biomass would be baled
and delivered as bales; (b) switchgrass has a nine month harvest
window and forage sorghum a five month harvest window; and
(c) forage sorghum requires twice as much time between mowing
and baling. For the region a harvest system other than baling may
be required for forage sorghum to be economically competitive
with switchgrass.
Acknowledgements
Support for this student training project was provided by USDA
National Needs Graduate Fellowship Competitive Grant No. 200838420-04777 from the National Institute of Food and Agriculture.
This project is also supported by the USDA National Institute of
Food and Agriculture, Hatch Grant Number H-2824, and the Oklahoma Agricultural Experiment Station. Support does not constitute
an endorsement of the views expressed in this paper by the USDA.
References
[1] Congress, U.S. Section 111, Subtitle A, Renewable Fuels, Consumer Protection,
and Energy Efficiency Act of 2007, H.R. 6 (EAS); 2007.
[2] U.S. Department of Energy. U.S. billion-ton update: Biomass supply for a
bioenergy and bioproducts industry. Oak Ridge National Laboratory, Oak
Ridge, Tennessee; 2011.
[3] Rogers JK, Motal FJ, Mosali J. Yield, yield distribution, and forage quality of
warm-season perennial grasses grown for pasture or biofuel in the Southern
Great Plains. ISRN Agron; 2012. p. 7.
[4] Khanna M, Dhungana B, Clifton-Brown J. Costs of producing miscanthus and
switchgrass for bioenergy in Illinois. Biomass Bioenergy 2008;32(6):482–93.
[5] Aravindhakshan SC, Epplin FM, Taliaferro CM. Economics of switchgrass and
miscanthus relative to coal as feedstock for generating electricity. Biomass
Bioenergy 2010;34(9):1375–83.
[6] McCutchen BF, Avant Jr. RV, Baltensperger D. High-tonnage dedicated energy
crops: the potential of sorghum and energy cane. NABC Report 20, Reshaping
American Agriculture to Meet its Biofuel and Biopolymer Roles. National
Agricultural Biotechnology Council Report; 2008. p. 119–22.
[7] Rooney WL, Blumenthal J, Bean B, Mullet JE. Designing sorghum as a dedicated
bioenergy feedstock. Biofuels, Bioproducts Biorefining 2007;1(2):147–57.
[8] Hallam A, Anderson IC, Buxton DR. Comparative economic analysis of
perennial, annual, and intercrops for biomass production. Biomass Bioenergy
2001;21(6):407–24.
[9] Duffy M, Nanhou VY. Costs of producing switchgrass for biomass in Southern
Iowa, PM 1866. Ames Iowa: Iowa State University Extension; 2001.
[10] Perrin R, Vogel K, Schmer M, Mitchell R. Farm-scale production cost of
switchgrass for biomass. Bioenergy Res 2008;1(1):91–7.
[11] Dumler TJ, Shoup D, Martin KL. Forage sorghum silage cost-return budget in
South Central Kansas, vol. MF-648, Kansas State University Agricultural
Experiment Station and Cooperative Extension Service, Manhattan Kansas;
2009.
54
A.P. Griffith et al. / Applied Energy 127 (2014) 44–54
[12] Colombini S, Rapetti L, Colombo D, Galassi G, Crovetto GM. Brown midrib
forage sorghum silage for the dairy cow: nutritive value and comparison with
corn silage in the diet. Ital J Anim Sci 2010;9(3):273–7.
[13] Kruse A, Krupka A, Schwarzkopf V, Gamard C, Henningsen T. Influence of
proteins on the hydrothermal gasification and liquefaction of biomass.
Comparison of different feedstocks. Ind Eng Chem Res 2005;44(9):3013–20.
[14] Thomason WE, Raun WR, Johnson GV, Taliaferro CM, Freeman KW, Wynn KJ,
et al. Switchgrass response to harvest frequency and time and rate of applied
nitrogen. J Plant Nutr 2004;27(7):1199–226.
[15] Weyerhaeuser. Low-cost, high value harvesting. <http://www.weyerhaeuser.
com/Businesses/Timberlands/OptimizingHarvests>; 2012 [accessed April 2012].
[16] Rosenthal P. Vertical integration in international ethanol chains. Master’s
thesis. Wageningen (Netherlands): Wageningen University; 2006.
[17] Chan JH, Reiner D. Dynamics of evolution in the global fuel-ethanol
industry. University of Cambridge; 2011.
[18] Carolan J, Joshi S, Dale B. Technical and financial feasibility analysis of
distributed bioprocessing using regional biomass pre-processing centers. J
Agric Food Ind Org 2007;5(2). Article 10..
[19] Kazi FK, Fortman JA, Anex RP, Hsu DD, Aden A, Dutta A, et al. Techno-economic
comparison of process technologies for biochemical ethanol production from
corn stover. Fuel 2010;89:S20–8.
[20] Wright MM, Daugaard DE, Satriob JA, Brown RC. Techno-economic comparison
of biomass-to-biofuels pathways. Fuel 2010;89(1):S2–S10.
[21] Wright M, Brown RC. Establishing the optimal sizes of different kinds of
biorefineries. Biofuels, Bioproducts Biorefining 2007;1:191–200.
[22] Reynolds JH, Walker DL, Kirchner MJ. Nitrogen removal in switchgrass biomass
under two harvest systems. Biomass Bioenergy 2000;19(5):281–6.
[23] Vogel KP, Brejda JJ, Walters DT, Buxton DR. Switchgrass biomass production in
the Midwest USA: harvest and nitrogen management. Agron J
2002;94(3):413–20.
[24] U.S. Department of Agriculture, Natural Resources Conservation Service.
Planting and managing switchgrass as a biomass energy crop. <http://
www.nrcs.usda.gov/Internet/FSE_DOCUMENTS/stelprdb1042293.pdf>; 2009
[accessed April 2012].
[25] U.S. Department of Agriculture. Census of Agriculture, vol. 1. Oklahoma State
and County Data. Geographic Area Series; 2002.
[26] Tolbert VR, Wright LL. Environmental enhancement of U.S. biomass crop
technologies research results to date. Biomass Bioenergy 1998;15(1):93–100.
[27] Data.gov. Conservation Reserve Program average payments by county.
<http://explore.data.gov/Agriculture/Conservation-Reserve-Program-AveragePayments-by-C/22vv-yvgf>; 2010 [accessed January 2012].
[28] Fewell J, Bergtold J, Williams J. Farmers’ willingness to grow switchgrass as a
cellulosic bioenergy crop: a stated choice approach. Selected paper presented
at the 2011 Joint Annual Meeting of the Canadian Agricultural Economics
Society & Western Agricultural Economics; 2011.
[29] U.S. Department of Agriculture, National Agricultural Statistical Service. Land
values and cash rents. <http://quickstats.nass.usda.gov/results/E3B17739DDB3-3A1F-9DB3-AE347CF7B9FA>; 2010 [accessed January 2012].
[30] Song F, Zhao J, Swinton SM. Switching to perennial energy crops under
uncertainty and costly reversibility. Am J Agric Econ 2011;93(3):764–83.
[31] Jager HI, Baskaran LM, Brandt CC, Davis EB, Gunderson C, Wullschleger S.
Empirical geographic modeling of switchgrass yields in the United States. Glob
Change Biol Bioenergy 2010;2(5):248–57.
[32] Haque M, Epplin FM, Taliaferro CM. Nitrogen and harvest frequency effect on
yield and cost for four perennial grasses. Agron J 2009;101(6):1463–9.
[33] Marsalis MA, Bean B. United sorghum checkoff program: Western forage
production guide. Lubbock, Texas; 2010.
[34] Butler T, Bean B. Forage sorghum production guide. <http://www.
soilcropandmore.info/crops/Annual-Forage-Sorghum/
ForageSorghumProdGuide.pdf> [accessed September 2012].
[35] Marsalis MA. Advantages of forage sorghum for silage in limited input
systems. Selected paper presented at the 2011 Western Alfalfa & Forage
Conference; 2011.
[36] Hwang S, Epplin FM, Lee B, Huhnke RL. A probabilistic estimate of the
frequency of mowing and baling days available in Oklahoma USA for the
harvest of switchgrass for use in biorefineries. Biomass Bioenergy
2009;33:1037–45.
[37] Thorsell SR, Epplin FM, Huhnke RL, Taliaferro CM. Economics of a coordinated
biorefinery feedstock harvest system: lignocellulosic biomass harvest cost.
Biomass Bioenergy 2004;22(4):327–37.
[38] English BC, Short C, Heady EO. The economic feasibility of crop residues as
auxiliary fuel in coal-fired power plants. Amer J Agric Econ
1981;63(4):636–44.
[39] Cundiff JS, Harris WL. Maximizing output-maximizing profits: production of
herbaceous biomass for fiber conversion should be a carefully managed
equipment-based enterprise. Resource 1995;2:8–9.
[40] Epplin FM. Cost to produce and deliver switchgrass biomass to an ethanolconversion facility in the Southern Plains of the United States. Biomass
Bioenergy 1996;11:459–67.
[41] Kumar A, Cameron JB, Flynn PC. Pipeline transport of biomass. Appl Biochem
Biotechnol 2004;113(1–3):27–39.
[42] Holmgren M, Casavant K, Jessup E, Eric. Strategic freight transportation
analysis review of transportation costs for alternative fuels. SFTA Research
Report # 25, Washington State University, Pullman Washington; 2007.
[43] Eksioglu SD, Acharya A, Leightley LE, Arora S. Analyzing the design and
management of biomass-to-biorefinery supply chain. Comput Ind Eng
2009;57:1342–52.
[44] Worley JW, Cundiff JS. Comparison of harvesting and transport issues when
biomass crops are handled as hay vs silage. Bioresour Technol 1996;
56(1):69–75.
[45] Sultana A, Kumar A. Optimal siting and size of bioenergy facilities using
geographic information system. Appl Energy 2012;94:192–201.
[46] Judd JD, Sarin SC, Cundiff JS. Design, modeling, and analysis of a feedstock
logistics system. Bioresour Technol 2012;103(1):209–18.
[47] Kumar A, Sokhansanj S. Switchgrass (Panicum Vigratum, L) delivery to a
biorefinery using integrated biomass supply analysis and logistics (IBSAL)
model. Bioresour Technol 2007;98(5):1033–44.
[48] Wang C. Economic analysis of delivering switchgrass to a biorefinery from
both the farmer’s and processor’s perspectives. Master’s thesis. Knoxville
(Tennessee): The University of Tennessee; 2009.
[49] Tembo GF, Epplin FM, Huhnke RL. Integrative investment appraisal of a
lignocellulosic biomass-to-ethanol industry. J Agric Res Econ 2003;
28(3):611–33.
[50] Mapemba LD, Epplin FM, Huhnke RL, Taliaferro CM. Herbaceous plant biomass
harvest and delivery cost with harvest segmented by month and number of
harvest machines endogenously determined. Biomass Bioenergy 2008;
32(11):1016–27.
[51] Daianova L, Dotzauer E, Thorin E, Yan J. Evaluation of a regional bioenergy
system with local production of biofuel for transportation, integrated with a
CHP plant. Appl Energy 2012;92:739–49.
[52] Zhang J, Osmani A, Awudu I, Gonela V. An integrated optimization model for
switchgrass-based bioethanol supply chain. Appl Energy 2012. <http://
dx.doi.org/10.1016/j.apenergy> 2012.06.054.
[53] American Petroleum Institute. State motor fuel taxes 2013. <http://
www.api.org/oil-and-natural-gas-overview/industry-economics/~/media/
Files/Statistics/state-motor-fuel-taxes-report.pdf>; 2013 [accessed August
2013].