Greenhouse Gas Emission of an Economically Optimized Switchgrass Supply Chain for Biofuel Production: A Case Study in Tennessee

Transcription

1 Greenhouse Gas Emission of an Economically Optimized Switchgrass Supply Chain for Biofuel Production: A Case Study in Tennessee Zidong Wang, T. Edward Yu, James A. Larson, and Burton C. English Department of Agricultural & Resource Economics University of Tennessee Selected Paper prepared for presentation at the Southern Agricultural Economics Association (SAEA) Annual Meeting, Orlando, Florida, 3-5 February 2013 Copyright 2013 by Wang, Yu, Larson and English. All rights reserved. Readers may make verbatim copies of this document for non commercial purposes by any means, provided that this copyright notice appears on all such copies. 1

2 Greenhouse Gas Emission of an Economically Optimized Switchgrass Supply Chain for Biofuel Production: A Case Study in Tennessee Abstract This study estimates the greenhouse gas (GHG) emissions of a switchgrass supply system for a potential biorefinery producing 50 million gallons of biofuel annually in Tennessee. Under the condition of minimizing the cost of feedstock, the estimated annual GHG emissions of the switchgrass supply system are 256 g CO 2 e/m 2. Key Words: Greenhouse gas emission; supply chain; optimization; switchgrass; biofuel JEL Classifications: Q16, Q51 1. Introduction 1.1 Background Bioenergy has gained increasing attention as an ecological alternative to fossil fuel for sustainable development. In 2005, a Renewable Fuel Standard (RFS) was first established by the U.S. Congress under the Energy Policy Act of 2005 to mandate at least 7.5 billion gallons of biofuel to be used for the national transportation fuel supply by This RFS was expanded to a new RFS2 in 2007 with the Energy Independence and Security Act of 2007 (EISA). The minimum biofuel usage mandated by RFS2 was 36 billion gallon by 2022 (EISA 2007). At least, 16 out of the 36 billion gallons of biofuel needed to be derived from lignocellulosic biomass (LCB), which had the potential to achieve 60% reduction in GHG emissions compared with the baseline lifecycle greenhouse gas emissions, determined as the life cycle GHG emission from gasoline or diesel in 2005 (EISA 2007). 2

3 The development of large-scale LCB supply chain, including biomass production, harvest, storage and transportation, could be significant for several reasons. LCB is a bulky feedstock considering its energy content, making it comparatively expensive to harvest, store and transport. LCB feedstock has long storage period and the weathering and precipitation during storage add the cost. Finally, the opportunity cost of converting other cropland into LCB production also needs to be considered. LCB supply chain also lead to GHG emissions. Land use change, energy consumption during switchgrass production and harvest, production of machinery and resources, and transportation are the major sources. Land use change can lead to GHG flux change as different crops have different carbon sequestration rates (Qin et al. 2012). In addition, different N fertilizer application rates lead to different N2O emissions of among crops such as corn, soybean and wheat, etc. (Del Grosso et al. 2001). The application of fertilizer and herbicide and utilization of farming machineries for LCB feedstock production consume energy and create GHG emissions. The production of those chemicals and equipment will all create GHG emissions as well. Moreover, transportation of LCB feedstock to the biorefinery is another source of GHG emissions. Switchgrass, a native perennial grass in North America, has been regarded as a potential LCB feedstock for biofuel production. In Tennessee, a state-funded program, Tennessee Biofuel Initiative (TBI) with a funding level of $70 million, was created in A total of 5,100 acres of switchgrass land and a pilot biorefinery were established (Jackson 2012). Moreover, a commercial-scale switchgrass-based biorefinery is under consideration by DuPont Danisco Cellulosic Ethanol in Tennessee (Brass, 2011). To support the switchgrass-based biofuel 3

4 industry, the development of its feedstock supply chain, its economic cost and GHG emissions need careful consideration. In this paper a case study across Tennessee was conducted to assess the GHG emissions of the switchgrass supply chain for a potential commercial-scale biorefinery plant. Assuming cost minimization as the objective of the biorefinery, the initial step is to find the optimal location, feedstock draw area and supply chain system for the biorefinery plant. The GHG emissions are then estimated based on the resource and energy usage of the cost-optimized supply chain. Factors such as soil type, available cropland and yield, road system and weather in Tennessee are considered. 1.2 GHG emissions of switchgrass supply chain Previous studies on GHG emissions of switchgrass supply chain vary from each other due to different assumptions, system boundaries and calculation logics. One major source of GHG emissions or offset from the switchgrass supply chain is land use change. Many studies showed that switchgrass had a high carbon sequestration rate, which leaded to a net GHG reduction when land use change is considered (Spatari et al. 2005). Other studies suggest that potential deforestation and other indirect land use changes actually lead increased GHG emissions (Searchinger et al. 2008). Moreover, Qin et al. (2011) argued that the effect of switchgrass carbon sequestration depended on the previous crop. They examined three major crop, corn, wheat and cotton. Conversion from wheat to switchgrass production showed highest annual carbon sequestration rate of 95 g CO 2 /m 2 while the conversion from soybeans only lead to 67 g CO 2 / m 2 per year. 4

5 N 2 O emissions from switchgrass production are another major GHG source. In some studies, N 2 O emissions were regarded as the largest source of GHG emissions of the switchgrass supply chain (e.g. Adler et al. 2007; Crutzen et al. 2008). However N 2 O emissions are also affected by land use change. Study has showed that switchgrass had low N application rate and high N use deficiency, which in turn contribute to the reduction of N 2 O emission compared with other crops (Monti et al. 2012). Results of the cradle-to-grave life cycle assessment of biofuel GHG emissions varied a lot from each other. A range from to 203 g CO 2 e/m 2 GHG emissions was observed from previous studies (see Table 1). As for the switchgrass supply chain emissions, two major sources are energy consumption during switchgrass production and harvest, and land use change. According to the studies of Kim and Dale (2004) and Spatari et al. (2005), the annual GHG emissions from switchgrass production and harvest ranged from 116 to 156 g CO 2 e/m 2. The contribution of land use change to total GHG emissions was of greater variance. Spatari et al. (2005) showed a net carbon sequestration rate of 1057 CO 2 e/m 2 in their study while studies from Searchinger et al. (2008) and Fargione et al. (2008) indicated net CO 2 emission. Emissions from switchgrass storage and transportation contribute to less GHG emissions compared with land use change or switchgrass production. Yu et al. (2011) has discussed the potential influence of switchgrass transportation on regional road system focusing on GHG and pollutant emissions. Indirect emissions from the production of farm machinery and materials such as fertilizer were also considered by many life cycle studies (e.g. Nelson et al. 2009; Wilson 2012). A list of previous studies about switchgrass supply chain emission was showed in Table Methods and Data 5

6 2.1 Optimal switchgrass supply chain of cost minimization The Bio-Energy Site and Technology Assessment model (BESTA) was used to determine the optimal switchgrass supply chain for potential large-scale production. BESTA is mixed integer programming model with the detailed structure available in Gao (2011). The objective is to minimize the total switchgrass chain cost bounded by land, labor and equipment availability, yield and feedstock demand. Factors such as monthly inventory and storage dry matter loss are considered in the model. In this study, the demand of switchgrass in Tennessee was from a potential biorefinery with an annual output of 50 million gallons of biofuel. Total 256 industrial park sites with access to water, power, and good transportation and storage systems were selected as candidate location for the assumed biorefinery plant. All the available cropland in Tennessee and within 50 miles of the state border was potential supply region for switchgrass. Double-crop land such as soybeanwheat land was not considered in this model since it was too costly to convert into switchgrass production. Two options, square bale and round bale, were tested for switchgrass harvest and storage. According to previous studies of switchgrass production and harvest in Tennessee (Larson et al. 2010), square bales stored with tarp and pallet would decrease the dry matter loss during storage and were more cost-efficient than square bales stored without tarp and pallet. On the contrary, similar benefit from tarp and pallet could not offset the cost of labor and material for round bales. In this study tarp and pallet options were included for square bale storage but not for round bale storage. The biorefinery plant was assumed to produce same amount of biofuel every month to utilize its facilities. About 1/3 of the total amount of harvested feedstocks were directly delivered into biorefinery during harvest season while the rest 2/3 were stored besides 6

7 the farmland and delivered to biorefinery from March to October. A list of the assumptions is in Table Greenhouse gas flux assessment The GHG emissions from switchgrass supply chain covers from switchgrass establishment, harvest, storage until transportation to the biorefinery. Emissions were categorized into four major components: emissions from land use change, emissions from energy consumption, secondary emissions from production of farm machinery, seed, fertilizer, pesticide for switchgrass production, and switchgrass transportation emissions Emissions from land use change The impact of land use change on GHG emissions is from change of carbon content in soil and underground biomass as well as the N 2 O emissions. A biogeochemistry model, DAYCENT, was employed to evaluate the emissions of specific soil property, land use change type and weather information. DAYCENT has been proved as a valid model to simulate soil C and N cycle with integrate factors by previous studies (e.g. Del Grosso et al. 2005; Li et al. 2006). Especially, DAYCENT has a better performance at predicting the relative differences with changes in input parameters (Chamberlain et al. 2011), making it useful comparing different land use conversions. Two scenarios were simulated and each of them has a time period of 60 years. In scenario 1, a certain type of crop such as corm, hay, etc. was planted and harvested for 60 years while in scenario 2 the land use change from crop to switchgrass occurred after 30 years. The difference of soil carbon change and N 2 O emission between these two scenarios was calculated. Annual weather data used for DAYCENT was acquired from DAYMET (available at: ). Focusing on the relative change among different type of land use 7

8 conversions, same annual weather data was applied for all the 60 years to avoid annual variance due to weather change. The soil property was based on the data from U.S. Geological Survey (available at: ). Eight sites among the four most popular soil texture types in Tennessee (loam, silt loam, silt clay loam, clay loam) were simulated with DAYCENT for each land use change type. The mean values of the simulated output including annual N 2 O emission and carbon sequestration rate were adopted for the assessment of emission (see Table 3) Emissions from energy consumption and indirect emissions Emission parameters from energy consumption during switchgrass production and harvest were adopted from well-cited life-cycle model GREET developed and maintained by the Argonne National Laboratory (available at: and were adjusted to fit the situation in Tennessee according to the Switchgrass Budget (UT extension, 2008). According to the Switchgrass Budget, tractor is the only source of energy consumption and other equipment such as baler, loader, mower and rake works together with tractor. The tractor used has a maximum PTO of 215HP, with the fuel usage of 9.42 gallon/hour. Based on the time consumption with each farming procedure, the energy consumption and emission parameters were calculated (see Table 3). Indirect emissions refer to GHG emitted during the production of agricultural machinery, fertilizer, herbicide and seed. The calculation of emissions from production of agricultural machinery was based on its weight and components following the logic of GREET. Weight information for equipment used in this study was estimated from the Official Guide: Tractor and Farm Equipment (Spring 2010) and producers such as John Deere. Fertilizer production 8

9 emission parameters were adopted from GREET model. The production emissions from the three kinds of pesticide used for switchgrass in Tennessee, Roundup, Cimarron and grass herbicide, were calculated based on Nelson et al. (2009). Emission parameters for switchgrass seed production was adopted from Wilson et al. (2012) Emissions from switchgrass transportation In this study, the Motor Vehicle Emissions Simulator (MOVES) was used to estimate the transportation emissions of switchgrass from field to biorefinery plant gate. On March 2, 2010, the MOVES model (MOVES2010) was firstly approved for official use outside of California by EPA. MOVES model is able to be adjusted for researches of different region and level and has its advantages to include impact factors such as travel speed, season, etc. into transportation emission assessment. The version used in this study is MOVES2010 (available at: Transportation emission inventories were created for the potential supply region and again aggregated at the county level to represent the base case conditions before the increase of traffic due to feedstock transportation. Next, the additional traffic due LCB feedstock transportation from farms to the potential biorefinery site estimated from BESTA model will be added to the base line case. The increase of GHG emissions represents those due to switchgrass transportation. 3. Results 9

10 3.1 Optimal switchgrass supply chain and cost BESTA model output showed that the square bale scenarios were more cost-effective than those round bale ones. The optimal candidate lied in Bedford County, middle Tennessee, which was also close to interstate I65 and I24. The total supply chain cost was $70 per ton of switchgrass delivered to the biorefinery plant gate. The total annual cost for the entire supply chain covering from land conversion until switchgrass delivery was $46 million for the optimal scenario in Middle Tennessee. Harvest cost accounted for 52% of the total cost, reaching $36 per ton. Switchgrass production and transportation both took up about 25% of the total cost while storage cost accounted for the rest 7% (see Fig. 1). 3.2 Land use change and resource consumption To meet the annual capacity of the potential biorefinery plant, about 79,920 acres of cropland in the nearby 10 counties were estimated to be converted into switchgrass production according to the BESTA output. More than 99% of the related switchgrass supply region was converted from previous hayland. The rest was converted from cotton, wheat, and soybean (see Fig. 2). Two reasons lead to large amount of hayland conversions: First of all hay is the major crop type in Middle Tennessee; Moreover hayland is less profitable than other crops such as corn, making it less costly for conversion. With the application rates of 3.4 and 27.2 kg per acre for switchgrass seed and nitrogen fertilizer relatively, about 272 tons of seed and 2174 tons of N fertilizer were used during switchgrass establishment and annual maintenance. Besides, 141 tons of Roundups, 6.4 tons of grass herbicide and 240 kg of Cimarron were used during switchgrass establishment and annual maintenance. 10

11 As for switchgrass harvest, a total of 741 balers, 1075 loaders, 288 mowers, 187 rakes and 2291 tractors were needed for switchgrass establishment and harvest according to estimation. A large number of tractors were used since they had to co-work with other agricultural equipment. Operations during switchgrass establishment, annual maintenance, harvest and storage lead to large energy consumption. Energy consumptions during switchgrass establishment (with annual maintenance included), harvest, and storage were 0.13, 1.58 and 0.04 million gallons respectively. As for switchgrass transportation, it took about 41,379 truckloads to deliver the switchgrass from field to biorefinery plant (see Table 4a). 3.3 GHG emission of switchgrass supply chain Based on the resource and energy input, the total GHG emitted from economically optimized feedstock supply chain system reached 82,767 CO 2 e ton (see Table 3c). Based on switchgrass delivered to the biorefinery gate, the GHG emission level is 126 CO 2 e kg per ton. Based on land converted, the average emissions were 256 g CO 2 e m -2 yr -1. Energy consumption during switchgrass establishment, harvest and storage was the largest GHG source, covering 44% of the total emissions. The distribution of the GHG emissions among these four procedures of switchgrass supply chain varied a lot due to the energy consumed in each step. Switchgrass harvest, the largest source of energy consumption, contributed to 39% of the total GHG emissions alone. Energy consumption during switchgrass production and storage contributed to 3% and 2% of the total GHG emission, respectively. Land use change also leaded to both CO 2 and N 2 O emissions, making up 31% of the GHG emissions. In detail, the N 2 O emissions from land use change leaded to 11% of the total GHG emissions. The CO 2 emissions from land use change covered 20% of the total GHG emissions. 11

12 Base on the type of land conversion, hay-to-switchgrass is the only source of net CO 2 emission of 16,732 CO 2 e ton. All the other types of conversion such as cotton-, soybean-, and wheat-toswitchgrass leaded to net carbon sequestration. However due to limited amount of land converted, these three types of land use change contributed to 158 ton of CO 2 sequestration in total. The MOVES output showed that switchgrass transportation leaded to 4,180 tonnages of CO 2 emissions, contributing to 5% percent of the total GHG emissions. Besides, output from MOVES also indicated that the CO 2 emissions from switchgrass transportation will increase the regional vehicle emission by 0.1% in the 10 counties where land use change to switchgrass production happened. The production emissions of machinery, fertilizer, seed and pesticide took 19% of the total GHG emissions. Specifically, the production of fertilizer and machinery took 10% and 7% respectively. Emissions from seed and pesticide production only covered less than 2% of the total emissions together. 3.4 Offset effect of N 2 O emission When assessing the CO 2 emissions from land use change, the difference of carbon sequestration between previous crop and switchgrass was considered to calculate CO 2 emission. On the other, the N 2 O emission the calculated based on the net emission from switchgrass production. This is the assumption adopted in this study and many other studies (e.g. Adler et al. 2007) and life cycle models such as GREET. However, difference of N 2 O emission between previous crop and switchgrass might also exist. LCB crops such as switchgrass had less fertilizer application rate than crops such as corn 12

13 (e.g. Lynd 1996). Since N fertilizer is the major source for soil N 2 O emissions, the land use change from other crops to switchgrass might need to less N 2 O emissions. Simulation output from DAYCENT showed similar results: The annual N 2 O emissions of switchgrass are significantly different from the previous crop type (see Fig.4). Specifically, except for soybean, which can fix nitrogen itself, the conversion from other crops to switchgrass reduced less N 2 O emission. Taking this effect into account, Soil N 2 O emission will change from a GHG source to a GHG sink since the major land conversion is hay-to-switchgrass. Applying this to our analysis of the optimal supply chain in middle Tennessee, we observed the soil N 2 O emissions change from 9,400 ton to -3,159 ton. The total GHG emission from the switchgrass supply chain thus reduced by 20%. 4. Conclusion and Discussion In this study a regional-scope GHG emissions assessment of LCB feedstock supply chain was introduced based on a pre-optimized feedstock supply chain with the objective of cost minimization. Study showed that the economically optimized switchgrass supply chain in Tennessee, with a plant-gate cost of $70 per ton, was a net GHG emission source with an annual emission rate of 256 g CO 2 e m -2 yr -1. Switchgrass production emissions including energy consumption of switchgrass production and harvest had the largest GHG emissions of 108 g CO 2 e/m 2. The land use change also leaded to 51 g/m 2 CO 2 emissions and 29 g CO 2 e/m 2 N 2 O emissions respectively since conversion of less profitable hayland lead to net soil carbon loss instead of sequestration. Moreover, if the difference of annual N 2 O emission rate among different crop is considered, land use change to switchgrass actually lead to net N 2 O emission 13

14 reduction in the Tennessee. This offset lead to a 20% reduction in total supply chain GHG emissions. The switchgrass production emission of 108 g CO 2 e/m 2 in this study was close to previous studies by Spatari et al. (2005) and Kim and Dale (2004) (see Table 1). Unlike many previous studies in which land use change leaded to net carbon sequestration, the net CO 2 emission of 51 g/m 2 CO 2 in this study resulted from the specific hay-to-switchgrass land use change occurred. Similar results indicating the net carbon loss from the change of other perennial species to switchgrass could be found in Davis et al. (2010) and Zan et al. (2001). The N 2 O emission level of 29 g CO 2 e/m 2 is lower than previous studies such as 40 g CO 2 e/m 2 (Adler et al. 2007) or 58 g CO 2 e/m 2 (Grandy et al. 2006). The no-till farming practice in Tennessee is the potential reason since the N 2 O emission for no-till crop is lower than till crop by DAYCENT simulation (Del Grosso et al. 2001). In this case, the GHG emissions are calculated based on the cost-minimized switchgrass supply chain. This is the reason why the major land conversion is the less costly hay-toswitchgrass and net carbon loss instead of sequestration occurred. And it is expected that lower GHG emissions and higher supply chain cost can be achieved if more land conversion were from crops such as corn and cotton. Since both economic cost and the environmental impact determine the future development of LCB-biofuel, it is necessary to balance these two objectives. Future studies with duel-objective optimization are needed to assess both the economic cost and GHG emission of the switchgrass supply chain as the feedstock for biofuel production. 14

15 Reference Adler, P. R., D. Grosso, S. J., and W. J. Parton. Life-cycle Assessment of Net Greenhouse-gas Fux for Bioenergy Cropping Systems. Ecological Applications 17: , Brass, L. East Tennessee Farmers First Switchgrass; Harvest Ready to Turn Into Biofuel. Available at: y_turn_biofuel Chamberlain. J. F., S. Miller, J. Frederick. Using DAYCENT to quantify on-farm GHG emissions and N dynamics of land use conversion to N-managed switchgrass in the Southern U.S. Agriculture, Ecosystems and Environment 141: , Cherubini, F., and G. Jungmeier. LCA of a Biorefinery Concept Producing Bioethanol, Bioenergy, and Chemicals from Switchgrass. International Journal of Life Cycle Assessment 15:53-66, Davis, S.C., W.J. Parton, and F.G. Dohleman. Comparative biogeochemical cycles of bioenergy crops reveal nitrogen-fixation and low greenhouse gas emissions in a Miscanthus xgiganteus agro-ecosystem. Ecosystems 13: , Del Grosso, S.J., W.J. Parton, A.R. Mosier, M.D. Hartman, J. Brenner, D.S. Ojima, and D.S. Schimel. Simulated interaction of carbon dynamics and nitrogen trace gas fluxes using the DAYCENT model. Modeling Carbon and Nitrogen Dynamics for Soil Management. CRC Press, Boca Raton, Florida, pp , Del Grosso, S.J., A.R. Mosier, W.J. Parton, and D.S. Ojima. DAYCENT model analysis of past and contemporary soil N2O and net greenhouse gas flux for major crops in the USA. Soil Tillage Research 83:9 24,

16 Energy Independence and Security Act of Public Law , Available at : Farrell, A.E., R.J., Plevin, B.T. Turner, A.D. Jones, M. O Hare, and D.M. Kammen. Ethanol can contribute to energy and environmental goals. Science 311: , Gao, Y. Evaluation of Pre-processing and Storage Options in Biomass Supply Logistics: A Case Study in East Tennessee. Master Thesis. University of Tennessee, Knoxville, Grandy, A., T. Loecke, S. Parr, and G. Robertson. Long-Term Trends in Nitrous Oxide Emissions, Soil Nitrogen, and Crop Yields of Till and No-Till Cropping Systems. Journal of Environmental Quality 35: , Jackson, S UTBI: An Update on Progress and Future Plans. Available at: Genera_Brief%20Overview_Keyser.pdf Kim, S., and B. E. Dale. Cumulative Energy and Global Warming Impact from the Production of Biomass for Biobased Products. Journal of Industrial Ecology 7: , Larson, J. A., D.F. Mooney, B. C. English, and D.D Tyler. Cost Analysis of Alternative Harvest and Storage Methods for Switchgrass in the Southeastern U.S. Paper presented at Southern Agricultural Economics Association Annual Meeting, Orlando, FL: February 6-9, Li, X., T. Meixner, J. O. Sickman, A. E. Miller, J. P. Schimel, and J. M. Melack. Decadal-scale Dynamics of Water, Carbon and Nitrogen in a California Chaparral Ecosystem: DAYCENT Modeling Results. Biogeochemistry 77:217:245, Lynd, L.R. Overview and Evaluation of Fuel Ethanol from Cellulosic Biomass: Technology, Economics, the Environment, and Policy. Annual Review of Energy and the Environment 21: ,

17 Nelson, R.G., C. M. Hellwinckel, C. C. Brandt, T. O. West, D. De La Torre Ugarte, G. Marland. Energy Use and Carbon Dioxide Emissions from Cropland Production in the United States, Journal of Environmental Quality 38: , Ney, R.A., and J.L. Schnoor. Greenhouse Gas Emission Impacts of Substituting Switchgrass for Coal in Electric Generation: The Chariton Valley Biomass Project. Center for Global and Regional Environmental Research. May 20, Pimentel, D., and Patzek, T. Ethanol Production Using Corn, Switchgrass, and Wood; Biodiesel Production Using Soybean and Sunflower. Natural Resource Research 14:65-76, Qin, Z., Q. Zhuang, and M. Chen. Impacts of Land Use Change due to Biofuel Crops on Carbon Balance, Bioenergy Production, and Agricultural yield, in the Conterminous United States. GCB Bioenergy 4(3): , Samson RA, Stamler BS. Going Green for Less: Cost-Effective Alternative Energy Sources. C.D. Howe Institute Commentary. ISSN , Toronto, Ontario, 28, Searchinger, T., R. Heimlich, R.A. Houghton, F. Dong, A. Elobeid, J. Fabiosa, S. Tokgoz, D. Hayes, and T.-H.Yu. Use of U.S. Croplands for Biofuels Increases Greenhouse Gases Through Emissions from Land-Use Change. Science February, pp , Spatari, S., Y. Zhang, and H.L. MacLean. Life Cycle Assessment of Switchgrass- and Corn Stover-Derived Ethanol-Fueled Automobiles. Environmental Science and Technology 39: , Thorsell, S., F. M. Epplin, R. L. Huhnke, and C. M. Taliaferro. Economics of a Coordinated Biorefinery Feedstock Harvest System: Lignocellulosic Biomass Harvest Cost. Biomass and Bioenergy 27(4): ,

18 University of Tennessee Extension. Guideline Switchgrass Establishment and Annual Production Budgets Over Three year Planning Horizon Available at: Vadas, P., K. Barnett, and D. Undersander. Economics and Energy of Ethanol Production from Alfalfa, Corn, and Switchgrass in the Upper Midwest, USA. Bioenergy Resource 1:44-55, Wilson, T.O., F. M. McNeal, S. Spatari, D. G. Abler, and P. Adler. Densified Biomass Can Cost-effectively Mitigate Greenhouse Gas Emissions and Address Energy Security in Thermal Applications. Environmental Science and Technology 46(2): , Wu, M., Y. Wu, and M. Wang. Energy and emission benefits of alternative transportation liquid fuels derived from switchgrass: a fuel life cycle assessment. Biotechnology Progress, 22: , You, F., and B. Wang. Life Cycle Optimization of Biomass-to-Liquid Supply Chains with Distributed-Centralized Processing Networks. Industrial and Chemistry Research 50: , Yu, T-H.E, S. Tokgoz, E. Wailes, and E. Chavez. A Quantitative Analysis of Trade Policy Responses to Higher World Agricultural Commodity prices. Food Policy 36: , Zan, C.S., J.W. Fyles, P. Girouard, and R.A. Samson. Carbon sequestration in perennial bioenergy, annual corn and uncultivated systems in southern Quebec. Agricultural Ecosystem Environ 86: ,

19 Table 1. Comparison of Annual CO 2 and N 2 O Emission Change Due to Land Use Change to Switchgrass Production Author and Year System boundary Final Product Value Unit Site (or land type) Model used Qin et al Carbon sequestration Biofuel g CO From soybean to e/m 2 Terrestrial switchgrass Ecosystem g CO From wheat to Model e/m 2 switchgrass Kim and Dale 2004 Switchgrass production Switchgrass 131 to 156 g CO 2 e/m 2 Seven states in U.S. Switchgrass production 116 Switchgrass production; direct -941 g CO Spatari et al land use change Ethanol 2 e/m 2 Ontario, Canada LCA Cradle to grave with direct 129 land use change Cherubini and Jungmeier 2010 Samson and Stamler 2009 Wu et al Cradle to grave with direct land use change Cradle to grave with direct land use change Cradle to grave with direct land use change Biofuel and other products 203 Ethanol -135 Ethanol and other products to -766 g CO 2 e/m 2 with set-aside land SimaPro 7 g CO 2 e/m 2 g CO 2 e/m 2 Ontario, Canada GREET, GHGenius GREET 19

20 Table 2. Supply Chain Operation Options and Assumptions 1. Biorefinery Feedstock: Switchgrass Capacity: 50 million gallon per year, 4.17 billion per month Requirement: Industrial parks with access to water, power, and roads, as well as sufficient storage space 2. Switchgrass Supply Chain Draw area: Cropland within 75 miles of the biorefinery (federal land and soybean-wheat double crop land excluded) Harvest season: November to February Harvest/storage: 1) Round bale, untarp, unpallet, 2) Square bale, tarp, pallet Transportation: 1) ton per truck for round bale 2) ton per truck for square bale Table 3. Emission Factors for Switchgrass Cropping System (Unit: CO2 e-kg/acre/year) Items Value Source Corn Cotton Land use change CO 2 Hay Sorghum DAYCENT Soybean Wheat Corn Cotton Land use change N 2 O Hay Sorghum DAYCENT Soybean Wheat Tractor Loader Farm and harvest machine Square baler Round baler GREET Mower PTO rake Production Energy consume Harvest GREET Storage 2.32* Fertilizer Production of fertilizer, seed Seed and herbicide Herbicide 1.34 GREET 20

21 Table 4. Resources Used in Switchgrass Cropping System and Associated Greenhouse Gas Emissions (a) Resources Item Unit Value Energy Production 0.13 Million gallon Harvest 1.58 diesel Storage 0.04 Equipment Baler 741 Loader 1075 Number of Mower 288 equipment Rake 187 Tractor 2291 Material Seed Tonnage 272 Fertilizer 2174 Pesticide 148* *: 141 tons are Roundup, 6.4 tons are grass herbicide and 0.2 tons are Cimarron. (b) Greenhouse Gas Emissions GHG emission Sources (CO 2 e ton) Land use CO 2 16,574 Direct change N 2 O 9,400 Production 2,653 Energy Harvest 32,381 Consumption Storage 1,523 Feedstock transportation 4,180 Machine production 5,823 Indirect Fertilizer 8,511 Herbicide 107 Seed 1,451 Total 82,603 Emission (g CO 2 /m 2 )

22 Supply chain cost ($/ton) Production Harvest Storage Transportation East (McMinn) Middle (Bedford) West (Weakley) Figure 1. Optimal Supply Chain Cost for the Rank 1 Candidate in Each Region 22

23 g CO2 /m2 Figure 2. Optimal Location and Feedstock Supply Region in Tennessee for Biorefinery with Annual Capacity of 50 Million Gallon CO2 emission increase N2O emission increase (50) (100) (150) corn cotton hay soybean wheat Previous Crop Type Figure 3. Annual CO 2 and N 2 O Emission Change Due to Land Conversion from Previous Crop into Switchgrass 23