Life Cycle Assessment of Potential Biojet Fuel Production in the United

Sep 29, 2011 - Impact of Alternative Fuels on Emissions Characteristics of a Gas Turbine Engine – Part 1: Gaseous and Particulate Matter Emissions. ...
0 downloads 9 Views 5MB Size
Download PDF
POLICY ANALYSIS pubs.acs.org/est

Life Cycle Assessment of Potential Biojet Fuel Production in the United States Datu B. Agusdinata,*,† Fu Zhao,‡ Klein Ileleji,§ and Dan DeLaurentis|| System-of-Systems Laboratory, College of Engineering, ‡School of Mechanical Engineering and Division of Environmental and Ecological Engineering, §School of Agricultural and Biological Engineering, and School of Aeronautics and Astronautics, Purdue University 701 West Stadium Avenue, West Lafayette, Indiana 47907, United States )

†

bS Supporting Information ABSTRACT: The objective of this paper is to reveal to what degree biobased jet fuels (biojet) can reduce greenhouse gas (GHG) emissions from the U.S. aviation sector. A model of the supply and demand chain of biojet involving farmers, biorefineries, airlines, and policymakers is developed by considering factors that drive the decisions of actors (i.e., decision-makers and stakeholders) in the life cycle stages. Two kinds of feedstock are considered: oil-producing feedstock (i.e., camelina and algae) and lignocellulosic biomass (i.e., corn stover, switchgrass, and short rotation woody crops). By factoring in farmer/feedstock producer and biorefinery profitability requirements and risk attitudes, land availability and suitability, as well as a time delay and technological learning factor, a more realistic estimate of the level of biojet supply and emissions reduction can be developed under different oil price assumptions. Factors that drive biojet GHG emissions and unit production costs from each feedstock are identified and quantified. Overall, this study finds that at likely adoption rates biojet alone would not be sufficient to achieve the aviation emissions reduction target. In 2050, under high oil price scenario assumption, GHG emissions can be reduced to a level ranging from 55 to 92%, with a median value of 74%, compared to the 2005 baseline level.

1. INTRODUCTION The aviation sector, facing mounting pressures to reduce its greenhouse gas (GHG) emissions, has begun to consider the use of biobased jet fuels (henceforth called biojet) as alternatives to fossil fuels (henceforth called petrojet).1,2 One of the goals that has been put forward is to achieve carbon neutral growth by 2020 and reduce the GHG emissions by 50% compared to the 2005 baseline level by 2050.3 In the United States, the aviation sector is responsible for about 11% of the total transportation GHG emissions.4 A recent study has shown that biojet on a unit energy basis can potentially lower GHG emissions by up to 85% when compared to petrojet.5 For biojet to achieve its GHG emission reduction potential, technical and economic hurdles must be overcome. Alternative jet fuels must have characteristics sufficiently similar to current petrojet regardless of the feedstock and refining process (i.e., be a “drop-in” fuel). However, the biojet produced by current refinery processes does not contain aromatic compounds, which account for up to 25% of petrojet by volume and are needed for proper lubrication and sealing.6 This, along with the requirement to meet fuel density specifications for aviation r 2011 American Chemical Society

fuel, requires that biojet be blended with petrojet. Currently, a 50% 50% blend by volume between biojet and petrojet fuel is the norm for meeting fuel property and performance specifications and is thus used in this paper as the upper threshold for blending.7 Although life cycle assessment (LCA) provides a sound basis to evaluate the overall environmental impacts (including GHG emissions) of biofuels,8 traditionally LCA studies have focused mostly on the environmental performance of technology options and largely left out the economic aspect of the system in question9 or at most include economic performance as a separate part.10 12 To reveal the achievable environmental benefits of emerging technologies such as biofuels, the economic motives of actors (i.e., decision-makers and stakeholders) involved along the life cycle stages have to be considered along with technical advances. There have been increasing interests and efforts on Received: September 7, 2010 Accepted: September 29, 2011 Revised: September 26, 2011 Published: September 29, 2011 9133

dx.doi.org/10.1021/es202148g | Environ. Sci. Technol. 2011, 45, 9133–9143

Environmental Science & Technology developing LCA methodology along this line (e.g., 13, 14), but to our best knowledge there has been no study conducted which focuses on the GHG emissions reduction potential of biojet. The objective of this paper is to reveal the extent to which biojet can reduce U.S. aviation GHG emissions through the consideration of the role and perspective of relevant actors. Decisions made by actors based on their motives, interests, and responses to incentives determine whether a policy objective can be achieved.15 This study looks at solely the GHG emissions within the U.S. domestic context with regard to the air transportation market and feedstock production. There are currently two biojet production technologies available: (1) hydrotreating/hydrocracking process which uses vegetable oil as the feedstock,16 and (2) gasification followed by Fischer Tropsch synthesis and syncrude upgrading, which uses lignocellulosic feedstocks17 (see Supporting Information, SI). The following feedstock options are considered in this study: (1) oil-producing feedstock: camelina (as representative of low-input oilseeds), and algae, and (2) lignocellulosic biomass: short rotation woody crops (SRWCs), corn stover (as representative of agriculture residue), and switchgrass (as representative of herbaceous energy crops). Out of concerns to the disruption on food production and effects of land-use change, soybean is excluded even though it is an important feedstock for biodiesel.18,19

2. METHODOLOGIES The model developed in this paper contains two major elements. The first element calculates the costs and GHG emissions associated with different life cycle stages for different feedstock and production pathways. The second element is a forecast model to determine the actual biojet supply and demand, and hence the life cycle GHG emissions of the U.S. aviation industry. The feedstock and process parameter values that are used as the baseline case in this study are specified in Tables S1 S9, SI. An uncertainty analysis is then conducted using the probability distribution specified for certain parameters (Tables S12 S17, SI). 2.1. Biojet Production Cost and GHG Emissions at Different Life Cycle Stages. Because this study focuses on system level

interaction among the actors involved, U.S. average data are used when developing the life cycle inventory. For feedstock considered, average yield is used over the regions suitable for growing the feedstock. For biorefinery, effects of location are not considered. That is, it is assumed that the biorefinery will require same capital and operating costs and have same conversion efficiency no matter where it is located. We use the data entries in Ecoinvent v.2.0 (especially the RER processes) as the bases for life cycle inventory development. This is based on the technology similarity between U.S. and EU. To better reflect U.S. scenarios, we modify the Ecoinvent unit processes (“U”) by replacing the original electricity input with U.S. power grid mix.20 Given the limited coverage of USLCI database, this has been a common approach adopted by researchers.21,22 System expansion method is adopted to account for coproducts in terms of revenues and emission credits. For feedstock production, the functional unit is defined as per ton of feedstock delivered to a biorefinery, as this is also the unit used by actors to make transactions. Similarly, for fuel production, the functional unit is defined as per gallon of biojet delivered. In this study, climate change is the only environmental impact considered and it is measured by kg

POLICY ANALYSIS

CO2 equivalent. The IPCC AR4 (2007) version is used to convert CO2, CH4, and N2O into CO2e23 for time horizon 100 year. Emissions of N2O from nitrogen fertilizers applied during cultivation are included. Using both direct and indirect emission factors from IPCC AR4, 1.325% of applied N in fertilizer is emitted as N2O nitrogen. To predict feedstock and biojet production cost over the time period considered, the current total cost is first broken down into subcategories such as land rent, fertilizer, capital equipment, fuel, and labor. Table S1 lists the models used to forecast price change. General inflation is assumed for other costs involved. Camelina. Although camelina (Camelina sativa) is not the most common oilseed grown in the U.S., it is considered here because camelina-based jet fuel has been actually produced, albeit in small quantity and tested using airplane of commercial airlines.7 Field tests suggest that seeding camelina requires no till or minimum till. Drilling lightly or broadcasting with roller harrow is appropriate for planting. Harvesting can be done using combining (direct or swathed) with 6/64 to 3/64 slotted screens, which are the same for harvesting alfalfa. Camelina yield in marginal land (i.e., Conservation Reserve Program, CRP, land) has not yet been documented, so is highly uncertain. So, based on expert consultation,24 it is assumed that camelina yield in marginal land falls between 50 and 70% of that observed in field tests.25,26 A study by Montana State University Agriculture Extension provides a detailed cost breakdown of camelina cultivation and those data are adopted here.27 Since agronomic and crop production improvement strategies are just beginning to be applied, it is expected that camelina yield could increase significantly over years. This increase is assumed to be achieved through scientific improvements instead of increased fertilizer usage.28 Although camelina meal contains high levels of Omega-3 fatty acids, it also has glucosinolate content which limits its use as animal feed.29 So, it is assumed that most of the camelina meal will be used as fuel to replace mill residue. As a conservative assumption, it will be sold at mill residue price of $20/ton.30 The farming and preprocessing (oil extraction) parameters for camelina are given in Tables S2 and S3. Algae. Currently, there are mainly two approaches to grow algae: open pond and photobioreactors. Photobioreactors offer higher productivity and less evaporative water loss, but require significantly larger capital investment and operating cost. A couple of recent studies suggest that biodiesel derived from algae cultivated in photobioreactors can have GHG emissions up to ten times higher than that from open-pond cultivation.10,31,32 Also, the production cost of algae lipid in the photobioreactors case is 60 100% higher than that in the case of open-pond production. Therefore, in this study it is assumed that open-pond systems will be used for algae cultivation. Lipid will be extracted through mechanical pressing and spent biomass will be digested for methane generation. The EPA report Renewable Fuel Standard Program (RFS2) Regulatory Impact Analysis lists the achievable algae yield, lipid content, lipid production cost, and GHG emissions for three cases: (1) a base case which corresponds to a reasonable but still challenging target for the near future (the year 2022); (2) an aggressive case which assumes identification of a strain with near optimal growth rates and lipid content; and (3) a maximum case which represents the near theoretical maximum based on photosynthetic efficiencies.10 These predictions, along with the material/energy input data, are adopted in this study. For the current scenario, data are extrapolated using algae yield 9134

dx.doi.org/10.1021/es202148g |Environ. Sci. Technol. 2011, 45, 9133–9143

Environmental Science & Technology and lipid content reported throughout the literature.33 The cultivation parameters for algae are given in Table S4. Corn Stover. To avoid soil erosion, it is assumed that up to 50% of corn stover can be collected for biojet production. It is also assumed that annual corn stover yield is the same as the grain yield. Thus, the future yield of corn stover can be estimated by linearly extrapolating corn grain yield data between 1985 and 2009.10 Removing corn stover will require replenishing the soil with nutrients. The amount of N/P/K required for every ton of corn stover removed is assumed to be constant. N2O credit due to corn stover removal is included by assuming 0.45% N content in corn stover and 1.25% N in N2O avoided per unit of N in stover removed. Material inputs and cost data are derived from an EPA report10 (all data are averaged for three different farm sizes. i.e., 200, 400, and 800 acres).The plant size in the EPA report is 4000 ton/day while the plant in this study is 2000 ton/day. This will affect the transportation cost from satellite storage to the plant, which is adjusted accordingly. In the base case, the farmer profit margin is assumed to be 10% of the production cost. It should be noted that although the on-farm product cost suggested by the EPA report is largely the same as reported in previous studies, there are significant differences reported on the cost of transportation (from farm edge to satellite storage and from satellite storage to plant) and storage.34 The total delivered cost in the EPA report is in agreement with some more recent publications thus is believed to be a more realistic estimate.35 The cultivation parameters for corn stover are given in Table S5. Switchgrass. According to Lee et al.,36 the average switchgrass yield in the U.S. CRP land varies between 0.85 and 3.6 ton/acre. For material inputs and cost details, a recent study conducted by Iowa State University Agriculture Extension is adopted.37 To reduce fertilizer usage, switchgrass will be harvested once a year. After establishment, the stands can produce for the next 10 years. It has been reported that switchgrass yield responds to N application rate linearly with a maximum yield achieved at 112 kg N/ha. It is assumed that the yield increase in the future will come from improved species, field management, and timing of fertilizer application so the N application rate can be kept the same. Also, although switchgrass yield does not respond to P and K, they are needed to maintain soil nutrient level. The amount of P and K needed is calculated based on a per ton residue removed basis. For storage, the Iowa State study cited a higher storage cost than in the EPA report ($16.67/ton vs $8.89/ton) mainly due to a higher building cost. For comparison purposes, the same EPA number is used as in the corn stover case. The cultivation parameters for switchgrass are given in Table S6. Short Rotation Woody Crops (SRWCs). As assumed in the case of switchgrass, CRP land or other low productivity cropland will be used for growing SRWCs. Since there is very little data, if anything, published on SRWC yield on CRP, the yield on natural forest is used as a surrogate, which ranges from 1 to 3.8 ton/ha.38 Based on the maximum observed annual yield, it is expected that the yield will be doubled in 2030.39 It is assumed that yield increase is the result of scientific improvements rather than increased fertilizer application. That is, the amount of fertilizer applied per acre will not change over the years. Compared with switchgrass, SRWCs require much less chemical inputs. The production cost is dominated by land rent, stumpage, and harvesting. The cultivation parameters for SRWCs are given in Table S7. Production Process. Hydrotreating/Hydrocracking Based Conversion. The developer of the hydrotreating/hydrocracking technology suggests that for a plant of 350 000 m3/year diesel

POLICY ANALYSIS

production capacity, the typical Inside Battery Limits (ISBL) erected capital cost estimates are between 60 and 80 million.40 It is assumed that a biojet plant of the same size will have similar ISBL erected cost. Total capital costs are usually assumed to be twice the ISBL capital costs.41 Recently, the DOE provided a $241 million loan guarantee to support the construction of a 137million gallon per year renewable diesel facility using the technology.42 The capital cost investment is estimated from this amount of capital. Because the considered plant has a lower capacity, an economies of scale factor of 0.8 is assumed.43 For diesel production, Kalnes et al. gives the range of yields for diesel and coproducts. The low end of diesel yield is adopted since biojet yield is lower than that of diesel.22 The biorefinery parameters for oil-producing feedstock are given in Table S8. Gasification and FT (Fischer Tropsch)-Synthesis. It is assumed that a plant size of 2000 ton/day is a good balance between the economies of scale and transportation cost at the current feedstock yield. Swanson et al. investigated the production cost of biofuel production via gasification and FT synthesis for the nth plant.44 The data for the case of low temperature fluidized bed gasifier is used here due to its higher technology readiness. For the same plant size, EPA report gives the production cost for the year 2022 based on improved plant availability, reduced capital cost, and higher fuel yield. It should be noted that diesel is the target product in both reports and is used as a surrogate for biojet. The biorefinery parameters for lignocellulosic feedstock are given in Table S9. 2.2. Forecast Model. To evaluate actors’ decisions and their impacts on achievable GHG emissions reduction, a forecast model is employed as shown in Figure 1. The model presents simplified biojet supply and demand logic influenced by actor’s decisions, regulatory and land constraints, as well as the cost, technology, and dynamics. The model is implemented using the Matlab software. Policymakers. From the policymaker’s point of view, although this study focuses on the U.S. context, U.S. domestic policies on climate change are partly conditioned to the results of international negotiations. The U.S., for instance, would not commit to a large GHG emission cut without similar commitments from all other major emitting countries.45 Policymakers drive the biojet demand by setting policy goals to reduce GHG emissions. Based on the carbon reduction goal, a “committed” emissions reduction trajectory is established. The discrepancy between the committed reduction trajectory and the emissions due to demand growth creates demand for biojets over time (Figure S2, SI). In the model, three oil and petrojet price scenarios are adopted. The scenarios were extrapolated from the U.S. Energy Information Administration data46 and are designated as LowOil, ReferenceOil, and HighOil (Figure S3, SI). Airlines. For airlines, the decision to use biojet is mainly influenced by the degree of savings in carbon costs. To influence this decision, policymakers have leverage on establishing a carbon marketplace to regulate the CO2 price. The earliest possible certification of biojet standard is assumed to be 2013.5 The aggregate U. S. airline fleet has shown a robust trend of improved fuel efficiency.47 The efficiency is captured by what is known as payload fuel energy efficiency (PFEE) (kg-km/MJoule), which measures payload (i.e., passenger and aircraft belly cargo) and flight range with respect to one unit energy of fuel.48 Air travel demand is commonly represented as revenue-passenger-kilometer (RPK), which equals the number of passengers multiplied by the flight distance, a counterpart 9135

dx.doi.org/10.1021/es202148g |Environ. Sci. Technol. 2011, 45, 9133–9143

Environmental Science & Technology

POLICY ANALYSIS

Figure 1. Flowchart for biojet demand and supply algorithm based on actors’ decision criteria, policy drivers, technological learning, time variable, and land constraint.

of the vehicle-miles-traveled (VMT) measure for road transport. A simple compound growth model of RPK is implemented in this study. The PFEE and RPK projections are used to estimate the overall jet fuel consumptions and carbon emissions (more is described in SI). Biorefineries. For biorefineries, the decision to build a refinery plant depends on whether such an undertaking provides viable return on investment. Thus, the net present value (NPV) and internal rate of return (IRR) are used as decision criteria. If the NPV and IRR exceed the threshold value and there is enough biojet demand, the biorefinery plant will be built, creating demand for feedstocks. The potential market share of biojet is modeled based on the logic that because all biojet fuels are supposed to be delivered as “drop in” fuels, the market share of each fuel source would be mainly determined by its relative unit production cost.49 When the supply of a particular feedstock is limited by factors such as land availability, its actual market share will be lower than its

potential. In this case, a land allocation approach similar to that of Smeets et al.50 is employed: the market share gap will be filled by other unconstrained and economically viable feedstocks, proportionally to their potential market share (eq S5, SI). Feedstock Producers/Farmers. Farmers/feedstock producers will satisfy the demand only if they get a certain profit margin from producing the feedstock. A threshold value of 10% is set, above which the feedstock will be produced, provided that there is demand and land available.51 Applied to the feedstock production costs, the profit margin in effect will determine the price that biorefineries would need to pay for the feedstock. To forecast feedstock cost, the cost at the plant gate is broken down into contributions from land rent, fertilizer/herbicide cost, farm labor cost, fuel cost, and other costs (e.g., farm machinery, seed, harvest, and transport). For algae, an open-pond system (raceway type) is assumed, which is deemed more economically viable than a photobioreactor system. 9136

dx.doi.org/10.1021/es202148g |Environ. Sci. Technol. 2011, 45, 9133–9143

Environmental Science & Technology

POLICY ANALYSIS

Figure 2. Comparison of the 2013 unit life cycle GHG emissions estimates and their range of uncertainties for different feedstocks. The unit emissions are broken down into major constituents.

Land Availability and Feedstock Land Competition. Another key factor influencing biojet supply is land availability for feedstock. To minimize the impact of biofuel production (including biojet) on the food supply, only abandoned agricultural land in the U.S. is considered.10 Due to its climatic and soil conditions and considering feedstock characteristics, the same patch of land can be shared among multiple feedstocks or is only suitable to a certain feedstock. The land suitability of the feedstock has been estimated by the U.S. Department of Energy for switchgrass and SRWCs52 and by the U.S. Department of Agriculture for camelina, which is based on natural habitat observation.53 A complete state by state distribution of land availability and suitability for the three feedstocks is given in Table S11. For corn stover, 75 million dry ton/year is now available and 169.7 million dry ton/year is available in the long term.54 A large-scale algal cultivation is subject to the availability of saline groundwater, solar radiation, and large stationary sources of CO2.55 The competition among feedstock for land (i.e., between camelina, switchgrass, and SRWCs) was dealt with based on

their profit-making potential, as implied by the market share equation (eq S5 SI). The competition also considers revenues from soil organic carbon (SOC) sequestration as an important new income source for farmers.56 A sequestration of 800 kg C/ha/year is assumed for switchgrass56 and 1860 kg C/ha/ year is assumed for SRWCs.57 There is no evidence for carbon sequestration for camelina in the literature. A carbon price of $27 per metric ton CO2 based on EPA’s projection is adopted.58 It should be noted that there is also competition for feedstock (thus land) to support production of biofuels used by different transportation sectors (e.g., ground transportation, aviation, ocean transportation). However, this is beyond the scope of this paper. About 11% of the total land available for biofuel production was allocated to biojet in this study. This fraction resembles the share of energy consumed by air transportation relative to the whole transportation sector.4 Experience/Learning Curve Effect. As the biojet production accumulates over time, the production costs will decline due to learning (i.e., experience curve). For the baseline case, a progress 9137

dx.doi.org/10.1021/es202148g |Environ. Sci. Technol. 2011, 45, 9133–9143

Environmental Science & Technology

POLICY ANALYSIS

Figure 3. Unit production cost of biojet comprising farming and refining components and the uncertainty range in 2013.

ratio of 81% for refinery cost reductions was assumed, which is derived from experiences within the bioethanol industry.59 In other words, when the accumulated biojet production doubles, the refinery costs will decline by 19%. Time Delay and Risk Attitude. Time delays create inertia in the biojet production between demand signal and actual supply.60 A time delay factors in the cultivation and biorefinery plant construction time. The length of time delay was modeled as a function of fuel price based on the assumption that the higher the fuel price, the more risks the actors are willing to take. This risk-seeking attitude will manifest by, for example, starting the plant construction earlier in anticipation of future demand.61 For the baseline case, a total of 3, 2, and 1 year delays are assumed for the LowOil, ReferenceOil, and HighOil scenarios, respectively. Furthermore, actors’ more risk-seeking attitude as the price of oil increases is reflected also in how they select the decision criteria threshold value. First, the IRR threshold is assumed to decrease from 15% in the LowOil to 10% and 7.5% in the ReferenceOil and High Oil, respectively. Second, in the LowOil scenario, a biorefinery plant will be built only if the biojet demand equals or exceeds the plant design capacity. This criterion is relaxed in the ReferenceOil and HighOil to become 0.9 and 0.75, respectively. In the sensitivity analysis, the value of these parameters is varied.

3. RESULTS AND ANALYSIS 3.1. GHG Emissions of Biojet. Figure 2 presents the estimated unit GHG emissions of biojet, in g CO2 e/MJoule, produced from the five feedstocks in 2013. The GHG emissions are broken down into eight components and for each of these categories, the uncertainty range from the baseline case is estimated (SI provides the specification of uncertain variables). Feedstock cultivation is largely responsible for GHG emissions, particularly in biojet from oil-containing feedstocks. The amount can reach as high as 90% of the total in the case of algae. For camelina, the use of fertilizer contributes about 70% of the total emissions. The land use effect due to soil organic carbon (SOC) sequestration (for switchgrass and SRWCs) has considerable contribution and uncertainty. Overall, all feedstocks have a lower unit of GHG emissions compared to the standard unit emission reference for petrojet, which is 85 g CO2e/MJ fuel.62 A sensitivity analysis on the emissions drivers confirms that yield is the most important factor influencing emissions for most feedstock (details in SI). The second most important factor is the efficiency of fertilizer usage by the plant (in kg feedstock/kg fertilizer). Moreover, the change in SOC sequestration has a more profound effect in switchgrass (9% reduction in emissions for every 1% SOC increase) than in SRWCs (3% reduction for every 1% SOC increase). 9138

dx.doi.org/10.1021/es202148g |Environ. Sci. Technol. 2011, 45, 9133–9143

Environmental Science & Technology

POLICY ANALYSIS

Figure 4. Comparison of biojet economic performance in the HighOil price scenario (baseline case). The drop in production cost is due to learning curve effect and consequently results in a rise in NPV and IRR.

3.2. Biojet Unit Production Costs. Figure 3 presents the breakdown of unit production cost per gallon of biojet for each feedstock in 2013. Camelina and corn stover have the lowest total unit cost whereas algae has the highest. The unit production cost is divided into farming and refining process to identify the contribution of each subprocess. The farming costs range between 45% (for corn stover) and 92% (for algae) of the total unit production costs. For camelina, the major cost driver is land costs. For algae, about 40% of the unit cost is due to capital costs. The analysis also shows that revenues obtained from refinery coproducts have a bigger influence than those from carbon sequestration in lowering the production costs. A sensitivity analysis performed illuminates some cost drivers. For all feedstocks, it is of no surprise that yield has the major influence in the unit cost. For camelina, for instance, a 1% increase in yield will reduce unit cost by almost 1%, whereas for switchgrass the reduction is about 0.4%. The second major cost driver is capital cost, which represents the capital needed for the acquisition of the equipment and machinery. Evolution of Biojet Economic Performance: Unit Production Cost, NPV, and IRR. To illustrate the economic performance of

feedstocks, Figure 4 shows the evolution of biojet unit production cost, net present value (NPV), and internal rate of return (IRR) of biojet in the HighOil price scenario (figures for the other scenarios are given in the SI). In terms of unit production cost, biojet from camelina and corn stover can compete with petrojet from the beginning of the time frame (2013) (Figure 4a). Switchgrass and SRWCs-derived biojet becomes competitive after 2020 whereas biojet from algae does so after 2040. By contrast, in the LowOil scenario, the unit production cost of all five feedstocks is higher than the price of petrojet, so none of them is viable for production (Figure S5, SI). The second criterion for biojet production is NPV (Figure 4b). In the HighOil scenario, the NPV figures are positive at year 2013 for all feedstocks except for SRWCs and algae, which will break even (NPV g 0) around 2025 and 2040, respectively. In the ReferenceOil scenario, only the production of biojet from camelina and corn stover is expected to result in a positive NPV region (Figure S6, SI). The threshold value for the IRR criterion is set at 15%, 10%, and 7.5% for the LowOil, ReferenceOil, and HighOil scenarios, respectively, serving as a higher investment barrier compared to 9139

dx.doi.org/10.1021/es202148g |Environ. Sci. Technol. 2011, 45, 9133–9143

Environmental Science & Technology

POLICY ANALYSIS

Figure 5. Emissions trajectories under the three oil price scenarios (baseline case). None of emissions trajectories achieve the 2050 reduction target.

the positive NPV criterion. In the HighOil scenario, there is a progression in time for feedstock to meet biorefinery’s criteria to be a viable biojet source. For example, camelina-derived biojet will become viable at year 2013 and algae by 2040. In the ReferenceOil scenario, camelina-based biojet is the only economically viable feedstock, which occurs around 2022. None of the feedstocks is viable in the LowOil scenario Consequently, biojet from a mix of feedstocks manifests at different time frame in the HighOil scenario. The supply mix constitutes corn stover, camelina, and switchgrass starting from 2022 onward when there is enough demand to justify a full industrial production (Figure S8c, SI). Biojet from SRWCs will be produced starting 2030 and algae-derived fuel after 2040. By contrast, in the ReferenceOil the biojet supply is dominated solely by camelina (Figure S8b, SI). There are several observations on the dynamics of system behavior. First, supply does not become available until 2022 in the HighOil scenario due to the demand and time delay. Second, the drop in production cost trend is caused mainly by the reduction in refining costs due to experience/learning rate effect. The decline in unit production costs results in a jump in the NPV and IRR figures. The algae-derived biojet unit price also benefits from camelina learning rate since both feedstocks share the same biorefinery facility. 3.3. GHG Emissions Trajectories. Figure 5 shows the emissions trajectories under the three oil price scenarios using the baseline case set up. As a business-as-usual case is the emission trajectory generated without biojet use and a 2% annual demand growth. Due to the continuous improvement in the aircraft payload fuel energy efficiency, PFEE, the emissions grow at an annual average rate of around 0.67% but the emissions in 2050 could be 28% higher than the 2005 baseline level, widely missing the target of 50% below the 2005 baseline level. With biojet options, the realized emissions level in 2050 is 71% of the 2005 baseline level for the HighOil scenario. In the LowOil and ReferenceOil scenario, the emissions level is 128% (i.e., the same as the case without biojet) and 124%, respectively. In the HighOil price scenario, emissions begin to rise around 2038 after downward trend due to the constraint of the 50 50 blend requirement with petrojet (i.e., maximum biojet demand equals 50% of the total jet fuel consumption). By contrast, in the

ReferenceOil scenario, the supply of camelina based biojet is limited and cannot keep pace with the growth of air travel demand. In all three scenarios, the 2050 reduction target cannot be achieved. The assessment of uncertainty in the GHG emissions estimate is presented using a box plot at a 5-year interval (Figure 6). The plot applies for HighOil scenario and is generated from the 10 000 data points sampled using the Latin-Hypercube sampling method.63 In 2050, the median (i.e., 50th percentile, Q2) of the emissions is about 74% of the 2005 baseline level. In this scenario, the lowest emission level attainable (i.e., minimum value of 55% of the 2005 baseline level) is still above the 2050 reduction target. The condition is much worse in the ReferenceOil and LowOil scenarios, in which the minimum emission level is 120% and 128% of the 2005 baseline level, respectively (Figure S9, SI). The sensitivity analysis on aviation GHG emissions shows the extent of factors’ influence in achieving the reduction target (Figure S12, SI). The most significant factor is the fraction land dedicated for biojet production, followed by oil price.

4. DISCUSSION A commercial biojet fuel production system in the U.S. does not yet exist and thus offers opportunities for policy makers to influence its evolution to achieve the desired impacts. This study reveals some insights and implications for policy design. First, feedstock viability is conditional on two major factors: oil price and land availability. The lignocellulosic biomass based biojet (i.e., corn stover, switchgrass, and short rotation woody crops) only becomes viable when the oil price is high (i.e., the HighOil scenario). In this condition, its supply potential is more than four times larger than that of oil-producing feedstock (i.e., camelina and algae). When the oil price is lower (i.e., the ReferenceOil scenario), camelina is viable but its supply is constrained by suitability of land on which it can grow. Because policy makers may not intend to favor certain feedstock prematurely, they will need to consider the likelihood of oil price evolution. Second, to avoid potential competition with food production, the study considers the use of marginal lands for cultivating camelina, switchgrass, and SRWCs. The low productivity of this type of land, which can reach as low as half of that of cropland, 9140

dx.doi.org/10.1021/es202148g |Environ. Sci. Technol. 2011, 45, 9133–9143

Environmental Science & Technology

POLICY ANALYSIS

Figure 6. Uncertainty range of the evolution of aviation GHG emissions under the HighOil price scenario, given in a box plot depicting the minimum, quartile's, and maximum value.

can be detrimental for feedstock viability. A policy aiming to improve the productivity through a development of special feedstock variety, for example, should therefore become a priority. Third, it is also evident that biojet alone is not sufficient to achieve the 2050 GHG emission reduction target. Consequently, other measures are needed, including a steeper improvement in the fuel efficiency of the U.S. aircraft fleet than the current trend shows and more fuel efficient operational procedures. This result confirms the current projections, in which biojet will be responsible for a large share of reduction. Lastly, the result shows that the 50 50 blend requirement can significantly hamper the attainment of the policy goal. The relaxation of this requirement will be enabled by improvements in areas such as development of additives for improving biojet density and improving aircraft fuel tank to prevent leakage due to lack of aromatics compounds in biojet. The scope of the work may underestimate the amount of potential biojet supply as well as feedstock mix and hence GHG emissions impact. This work focuses only on U.S. production capacity and is limited to the consideration of marginal lands. It therefore excludes supply that may come from countries such as Canada and Brazil and a possibility that farmers may actually cultivate crop lands, resulting in higher yields. First generation feedstocks such as soybean cannot be ruled out completely. Due to its importance, different oil price scenarios such as price spikes and oscillations may result in different actor decision behaviors. Further work will need to address these issues. It should also include a further study of different incentive schemes that can be targeted to actors and feedstocks. The methodology presented in this paper can inform the design of incentives that are more aligned with actors’ interest. Also, the questions of “who should pay” and “how the costs and payoffs should be shared” encapsulate the central policy problem. This equity issue needs to be addressed with vigor commensurate with the technical evaluation (e.g., 64).

’ ASSOCIATED CONTENT

bS

Supporting Information. Parameter specifications and additional model descriptions and results. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*Phone: (765) 494-0418; fax: (765) 494-0307; e-mail: bagusdin@ purdue.edu.

’ REFERENCES (1) Kim, B. Y.; Fleming, G. G.; Lee, J. J.; Waitz, I. A.; Clarke, J. P.; Balasubramanian, S.; Malwitz, A.; Klima, K.; Locke, M.; Holsclaw, C. A.; Maurice, L. Q.; Gupta, M. L. System for assessing aviation’s global emissions (SAGE), Part 1: Model description and inventory results. Transp. Res., Part D-Transp. Environ. 2007, 12 (5), 325–346. (2) ATAG. Beginner’s Guide to Aviation Biofuels; Air Transport Action Group (ATAG), 2009. http://www.enviro.aero/Biofuels.aspx [accessed February 5, 2010]. (3) Butterworth-Hayes, P. Environmental regulations fly high and wide. Aerosp. Am. 2010, 4–6. (4) FAA. Aviation & Emissions: A Primer; Federal Aviation Administration, Office of Environment and Energy: Washington, DC, 2005. (5) Hileman, J. I.; Wong, H. M.; Ortiz, D.; Brown, N.; Maurice, L.; Rumizen, M., The feasibility and potential environmental benefits of alternative fuels for commercial aviation. In 26th International Congress of the Aeronautical Sciences (ICAS), Anchorage, AK, 2008. (6) Chevron. Aviation Fuels Technical Review; 2006. (7) Rahmes, T.; Kinder, J.; Crenfeldt, G. Sustainable Bio-Derived Synthetic Paraffinic Kerosene (Bio-SPK) Jet Fuel Flights and Engine Tests Program Results. In 9th AIAA Aviation Technology, Integration, and Operations Conference (ATIO) and Aircraft Noise and Emissions Reduction Symposium (ANERS), Hilton Head, SC, 2009. (8) Huo, H.; Wang, M.; Bloyd, C.; Putsche, V. Life-Cycle Assessment of Energy and Greenhouse Gas Effects of Soybean-Derived Biodiesel and Renewable Fuels; Technical Report, DOE ANL/ESD/08-2, Argonne National Laboratory; 2008. (9) Maclean, H.; Lave, L. Life cycle assessment of automobile/fuel options. Environ. Sci. Technol. 2003, 37 (23), 5445–5452. (10) U.S. EPA. Renewable Fuel Standard Program (RFS2) Regulatory Impact Analysis; Washington, DC, 2010. (11) Kou, N. N.; Zhao, F. Effect of multiple-feedstock strategy on the economic and environmental performance of thermochemical ethanol production under extreme weather conditions. Biomass Bioenergy 2011, 35 (1), 608–616. (12) Hang, Y.; Qu, M.; Zhao, F. Economical and environmental assessment of an optimized solar cooling system for a medium-sized benchmark office building in Los Angeles, California. Renewable Energy 2011, 36 (2), 648–658. 9141

dx.doi.org/10.1021/es202148g |Environ. Sci. Technol. 2011, 45, 9133–9143

Environmental Science & Technology (13) Grimes-Casey, H. G.; Seager, T. P.; Theise, T. L.; Powers, S. E. A game theory framework for cooperative management of refillable and disposable bottle lifecycles. J. Cleaner Prod. 2007, 15 (17), 1618–1627. (14) Davis, C.; Nikolic, I.; Dijkema, G. P. J. Integration of Life Cycle Assessment Into Agent-Based Modeling. J. Ind. Ecol. 2009, 13 (2), 306–325. (15) Walker, W. E. Policy Analysis: A Systematic Approach to Supporting Policymaking in the Public Sector. J. Multi-Criteria Decis. Anal. 2000, 9, 11–27. (16) Kinder, J. D.; Rahmes, T. Evaluation of Bio-Derived Synthetic Paraffinic Kerosene (Bio-SPK); The Boeing Company Sustainable Biofuels Research & Technology Program, 2009. (17) IATA. 2nd Generation Biomass Conversion Efficiency; International Air Transport Association: Montreal, 2010. (18) Hileman, J. I.; Ortiz, D. S.; Bartis, J. T.; Wong, H. M.; Donohoo, P. E.; Weiss, M. A.; Waitz, I. A. Near-Term Feasibility of Alternative Jet Fuels, Technical Report; Rand & Partner, 2009. (19) Williams, P. R. D.; Inman, D.; Aden, A.; Heath, G. A. Environmental and Sustainability Factors Associated With Next-Generation Biofuels in the US: What Do We Really Know? Environ. Sci. Technol. 2009, 43 (13), 4763–4775. (20) Ecoinvent. www.ecoinvent.org. (21) Hsu, D. D.; Inman, D.; Heath, G. A.; Wolfrum, E. J.; Mann, M. K.; Aden, A. Life Cycle Environmental Impacts of Selected US Ethanol Production and Use Pathways in 2022. Environ. Sci. Technol. 2010, 44 (13), 5289–5297. (22) Kalnes, T. N.; Koers, K. P.; Marker, T.; Shonnard, D. R. A Technoeconomic and Environmental Life Cycle Comparison of Green Diesel to Biodiesel and Syndiesel. Environ. Prog. Sustain. Energy 2009, 28 (1), 111–120. (23) Intergovernmental Panel on Climate Change (IPCC). Climate Change 2007: Synthesis Report; Geneva, 2007. (24) Johnson, K., Department of Agronomy, Purdue University, 2011. (25) McVay, K. A.; Lamb, P. F. Camelina Production in Montana; Montana State University Extension: Bozeman, MT, March 2008. (26) Ehrensing, D. T.; Guy, S. O. Camelina; Oregon State University Extension Service, 2008. (27) Johnson, D. Crops for Biodiesel Research in Montana: Camelina; Montana Department of Environmental Quality, Biodiesel Production Education Workshop: Livingston, MT, 2007. (28) Shonnard, D. R.; Williams, L.; Kalnes, T. N. Camelina-Derived Jet Fuel and Diesel: Sustainable Advanced Biofuels. Environ. Prog. Sustain. Energy 2010, 29 (3), 382–392. (29) Pekel, A. Y.; Patterson, P. H.; Hulet, R. M.; Acar, N.; Cravener, T. L.; Dowler, D. B.; Hunter, J. M. Dietary camelina meal versus flaxseed with and without supplemental copper for broiler chickens: Live performance and processing yield. Poultry Sci. 2009, 88 (11), 2392– 2398. (30) U. S. Environmental Protection Agency. Combined Heat and Power Partnership, Biomass Combined Heat and Power Catalog of Technologies; Washington, DC, 2007. (31) Jorquera, O.; Kiperstok, A.; Sales, E. A.; Embirucu, M.; Ghirardi, M. L. Comparative energy life-cycle analyses of microalgal biomass production in open ponds and photobioreactors. Bioresour. Technol. 2010, 101 (4), 1406–1413. (32) Stephenson, A. L.; Kazamia, E.; Dennis, J. S.; Howe, C. J.; Scott, S. A.; Smith, A. G. Life-Cycle Assessment of Potential Algal Biodiesel Production in the United Kingdom: A Comparison of Raceways and AirLift Tubular Bioreactors. Energy Fuels 2010, 24, 4062–4077. (33) Sheehan, J.; Dunahay, T.; Benemann, J.; Roessler, P. A Look Back at the U.S. Department of Energy’s Aquatic Species Program— Biodiesel from Algae; NREL/TP-580-24190; 1998. (34) Brechbill, S. C.; Tyner, W. E.; Ileleji, K. E. The Economics of Biomass Collection and Transportation and Its Supply to Indiana Cellulosic and Electric Utility Facilities. Bioenergy Res. 2011, 4 (2), 141–152. (35) Kazi, F. K.; Fortman, J. A.; Anex, R. P.; Hsu, D. D.; Aden, A.; Dutta, A.; Kothandaraman, G. Techno-economic comparison of process

POLICY ANALYSIS

technologies for biochemical ethanol production from corn stover. Fuel 2010, 89, S20–S28. (36) Lee, D. K.; Owens, V. N.; Doolittle, J. J. Switchgrass and soil carbon sequestration response to ammonium nitrate, manure, and harvest frequency on conservation reserve program land. Agron. J. 2007, 99 (2), 462–468. (37) ISU Ag Extension. Estimated Costs for Production, Storage and Transportation of Switchgrass; Iowa State University: Ames, IA, 2007. (38) Wright, L. L.; Hughes, E. E. United-States Carbon Offset Potential Using Biomass Energy-Systems. Water, Air Soil Pollut. 1993, 70 (1 4), 483–497. (39) Mann, M. K.; Spath, P. L. Life Cycle Assessment of a Biomass Gasification Combined-Cycle System; DOE TP-430-23076; Washington, DC, 1997. (40) Marker, T. L. Opportunities for Biorenewables in Oil Refineries; Final Technical Report, DOE GO15085, 2005. (41) Toman, M. J. G.; Lempert R. J. Impacts on U.S. Energy Expenditures and Greenhouse-Gas Emissions of Increasing RenewableEnergy Use; Technical Report TR-384-1; RAND Corporation, 2008. (42) DOE. Department of Energy Offers First Conditional Commitment for a Loan Guarantee for Advanced Biofuels Plant, 2011. http://www.energy.gov/9991.htm [accessed January 2011]. (43) Agusdinata, D. B. Exploratory Analysis to Support Real Options Analysis: An Example from Electricity Infrastructure Investment. IEEE International Conference on Systems, Man and Cybernetics, Waikoloa, Hawaii, 2005; Waikoloa, Hawaii, 2005. (44) Swanson, R. M.; Satrio, J. A.; Brown, R. C.; Platon, A.; Hsu, D. D. Techno-Economic Analysis of Biofuels Production Based on Gasification; NREL/TP-6A20-46587; 2010. (45) Committee on America's Climate Choices, National Research Council, America’s Climate Choices; The National Academies Press: Washington, DC, 2011. (46) EIA. Annual Energy Outlook 2010; U.S. Energy Information Administration: Washington, DC, 2010. (47) Lee, J.; Lukachko, S.; Waitz, I.; Schafer, A. Historical and future trends in aircraft performance, cost, and emissions. Ann. Rev. Energy Environ. 2001, 26, 167–200. (48) Hileman, J. I.; Katz, J. B.; Mantilla, J. G.; Fleming, G., Payload Fuel Energy Efficiency as a Metric for Aviation Environmental Performance. In 26th International Congress of the Aeronautical Sciences, Anchorage, AK, 2008. (49) Boyd, D. W.; Phillips, R. L.; Regulinski, S. G. A Model of Technology Selection by Cost Minimizing Producers. Manage. Sci. 1982, 28 (4), 418–424. (50) Smeets, E. M. W.; Faaij, A. P. C.; Lewandowski, I. M.; Turkenburg, W. C. A bottom-up assessment and review of global bio-energy potentials to 2050. Prog. Energy Combust. Sci. 2007, 33 (1), 56–106. (51) USDA. Structure and Finances of U.S. Farms: 2005 Family Farm Report; Economic Research Service-USDA: Washington, DC, 2005. (52) DOE. Breaking the Biological Barriers to Cellulosic Ethanol: A Joint Research Agenda. A Research Roadmap Resulting from the Biomass to Biofuels Workshop Sponsored by the U.S. Department of Energy, 2005. (53) USDA. http://plants.usda.gov/. (54) Perlack, R. D. W.; Turhollow, L. L.; Anthony, F.; Graham, R. L.; Stokes, B. J.; Erbach, D. C. Biomass As Feedstock for a Bioenergy and Bioproducts Industry: The Technical Feasibility of a Billion-Ton Annual Supply; U.S. Department of Energy and U.S. Department of Agriculture: Washington, DC, 2005. (55) Pienkos, P. T.; Darzins, A. The promise and challenges of microalgal-derived biofuels. Biofuels, Bioprod. Biorefin. 2009, 3 (4), 431–440. (56) Lemus, R.; Lal, R. Bioenergy crops and carbon sequestration. Crit. Rev. Plant Sci. 2005, 24 (1), 1–21. (57) Updegraff, K.; Baughman, M. J.; Taff, S. J. Environmental benefits of cropland conversion to hybrid poplar: Economic and policy considerations. Biomass Bioenergy 2004, 27 (5), 411–428. (58) U.S. Environmental Protection Agency. EPA Analysis of the American Clean Energy and Security Act of 2009 H.R. 2454 in the 111th 9142

dx.doi.org/10.1021/es202148g |Environ. Sci. Technol. 2011, 45, 9133–9143

Environmental Science & Technology

POLICY ANALYSIS

Congress, June 23, 2009. http://www.epa.gov/climatechange/economics/pdfs/HR2454_Analysis.pdf. (59) Bake, J. D. V.; Junginger, M.; Faaij, A.; Poot, T.; Walter, A. Explaining the experience curve: Cost reductions of Brazilian ethanol from sugarcane. Biomass Bioenergy 2009, 33 (4), 644–658. (60) Sterman, J. D. Business Dynamics: Systems Thinking and Modeling for a Complex World; Irwin/McGraw-Hill, 2000. (61) Neuhoff, K. Investment Decisions under Climate Policy Uncertainty; Electricity Policy Research Group, University of Cambridge, 2007. (62) Wong, H. M. Life-cycle Assessment of Greenhouse Gas Emissions from Alternative Jet Fuels; MIT, 2008. (63) Helton, J. C.; Davis, F. J. Latin hypercube sampling and the propagation of uncertainty in analyses of complex systems. Reliab. Eng. Syst. Saf. 2003, 81 (1), 23–69. (64) Agusdinata, D. B.; Delaurentis, D. A. Addressing equity issue in multi-actor policymaking via a system-of-systems approach: Aviation emissions reduction case study. J. Syst. Sci. Syst. Eng. 2011, 20 (1), 1–24.

9143

dx.doi.org/10.1021/es202148g |Environ. Sci. Technol. 2011, 45, 9133–9143