Challenges to the development of a dedicated energy crop.

The Energy Policy Act of 2005 includes a provision designed to double the production and use of ethanol in fuels by 2012, and to ensure that by 2013, a minimum of 250 million gallons per year of ethanol is produced from cellulosic sources, such as corn stover, wheat straw, and switchgrass. Cellulosic ethanol (CE) has a much greater potential volume than grain-based ethanol. For example, if the entire 2005 U.S. corn crop of 11.1 billion bushels had been converted to ethanol, the resulting product would have contained less than 9% of the energy contained in the 2005 U.S. net crude oil imports. Perlack et al. (2005) have estimated that it is technically feasible for the United States to produce more than a billion tons annually of cellulosic biomass that could be used as feedstock. If cellulosic biomass could be converted into ethanol at a rate of 90 gallons per dry ton, a billion tons could be used to produce ethanol containing approximately 26% of the BTUs of the 2005 U.S. net crude oil imports.

While some biomass could be obtained from wood wastes and crop residues, an energy crop will be required to obtain a billion tons annually. After extensive research, the Bioenergy Feedstock Development Program at the Oak Ridge National Laboratory selected switchgrass as a model biomass feedstock (McLaughlin et al. 1999; Fuentes and Taliaferro 2002). Perlack et al. (2005) posit that 55 million acres of cropland, idle cropland, and cropland pasture could be seeded to a dedicated perennial energy crop, such as switchgrass. Similarly, English et al. (2006) conclude that with some economic incentives, switchgrass could be established on more than 100 million U.S. acres.

Research and development is ongoing in an attempt to develop economically competitive methods to produce CE (Aden et al. 2002; McKendry 2002; Mosier et al. 2005; Service 2007; Wyman 1994). The U.S. Department of Energy's National Renewable Energy Laboratory (NREL) has established a goal of producing CE for $1.07 per gallon ($0.10 for enzymes, $0.58 for conversion, and $0.39 for feedstock) by 2012 (Pacheco 2006). NREL estimates that a yield of 90 gallons per dry ton and a delivered feedstock cost of $35 per dry ton could be attained.

If and when an economically competitive CE system is developed, it is anticipated that the agricultural community would be actively engaged in the production, harvest, storage, and transportation of feedstock to biorefineries. Relative to corn grain, cellulosic material, such as switchgrass is bulky and difficult to transport. Feedstock acquisition logistics for corn grain to ethanol plants are relatively simple--post a competitive price, and corn grain will be delivered by the existing marketing system. The infrastructure for production, harvest, storage, transportation, and price risk management of corn grain is well-developed; for switchgrass, it is virtually nonexistent.

In the absence of spot markets, obtaining a reliable flow of feedstock could involve: (a) contracting with individual growers; (b) contracting with a group of growers through a cooperative arrangement; (c) arranging long-term land leases similar to Conservation Reserve Program (CRP) leases; and/or (d) acquiring land. While land ownership may seem unreasonable, evidence shows otherwise. For example, one of the six companies awarded a U.S. Department of Energy contract to build a scaled-up CE facility expects to use feedstock produced exclusively on the more than 130,000 acres owned by the company.

Several potential incentive structures exist to achieve conversion of 50-100 million acres of U.S. land to the production of switchgrass as a dedicated energy crop. This article is focused on two of these structures. The first assumes a fully vertically integrated system, in which the biorefinery enters into long-term leases of land (similar to CRP leases). Production, harvest, storage, and transportation would be managed by the biorefinery. The second alternative assumes that the biorefinery enters into long-term production and harvest contracts with individual farmers. The farmers would be responsible for switchgrass establishment, management, and harvest, while the biorefinery would be responsible for transporting harvested feedstock from the farm to the biorefinery. The objectives of the research are: to determine the cost to produce switchgrass for both the land-lease alternative and the farmer-contract alternative, to determine if NREL's estimated delivered cost of $35 per dry ton is realistic, and to identify likely challenges to the development of switchgrass as a dedicated energy crop.

For the land-lease alternative, results of a model that includes 55 Oklahoma counties as potential production regions are presented. For the contract alternative, the process used to establish production contracts with Tennessee farmers is described and results are presented.

Methods--Land Leases

A multi-region, multi-period, mixed integer mathematical programming model similar to that described by Tembo, Epplin, and Huhnke (2003) and Mapemba et al. (2007) was constructed. The model was formulated and solved to determine the cost to produce, harvest, store, and transport a flow of switchgrass biomass to a biorefinery and identify the optimal biorefinery location from among several potential sites.

Expected yields were obtained from Graham, Allison, and Becker (1996) and Fuentes and Taliaferro (2002). Expected biomass yields differ across months of the year due to stage of growth and field losses that occur after plant maturation. Biorefinery size was based on biomass feedstock requirements of 2,000 dry tons per day. Based on estimates provided by forage storage specialists, storage losses were assumed to be 1% per month. The biorefinery was assumed to operate 350 days per year.

Assumptions regarding the type of harvesting method were based on modifications of results of a biomass harvest cost study reported by Thorsell et al. (2004). Based on their findings, it was assumed that biomass would be harvested in large, rectangular solid bales, stored in or near the production fields, and transported by truck to the biorefinery when needed. It was also assumed that harvest crews, either independent operators, or crews coordinated by central management of the biorefinery would conduct the harvest. Weather data were used to determine probability distributions for the number of days per month suitable for harvesting switchgrass for each county. The 95 % probability level from the harvest day distributions was selected so that the number of harvest days per month was set equal to the number of days that would be suitable for harvest in nineteen of twenty years. The model endogenously determines the number of harvest machines.

Shipment and processing of biomass can be done in any of the twelve discrete periods (months of the year). In months when biomass is harvested, it may be placed in storage or transported directly from the field to the biorefinery. Another assumption was that up to 10% of the cropland acres in a county could be bid away from other uses at a long-term lease rate of $60 per acre per year. The average 2006-7 cropland cash rental rate for Oklahoma dryland crop acres was $30 per acre, with a range from $10-$60 per acre (Doye and Sahs 2007). The $60 level was selected to enable management to lease land suitable for seeding to switchgrass production and to compensate land owners for a longer time commitment than required by an annual cash lease.

Two harvest systems were modeled. The first harvest season extended from July through February of the following year (eight-month system), while the second was restricted to July and August (two-month system). This restriction was imposed to determine how the length of the harvest season affects the cost of delivering a ton of switchgrass.

Methods--Production Contracts

The University of Tennessee Institute of Agriculture, through the federally funded Tennessee Switchgrass Project, was selected to determine the potential for large-scale production of switchgrass as a dedicated energy crop. One objective of the Tennessee project was to determine the incentive required to entice farmers to produce switchgrass. To achieve this objective, a competitive bidding process was conducted in the spring of 2005. The bidding process explicitly allowed the University to consider the bidder's ability to produce switchgrass and to evaluate their suitability for the project in addition to the amount bid. The factors used to determine bidder ability and suitability included: geographical location, experience, access to necessary equipment, cropping history of plot, and the range of acres the bidder was willing to devote to the program (Epplin et al. 2007).

The bids had two cost components: an annual base payment stated in dollars per acre and an incentive payment stated in dollars per ton of switchgrass produced per year. To evaluate the bids, a single per-acre total bid was calculated by assuming an average annual yield of 5.5 tons per acre. Thus, total per-acre bids were assumed to be equal to the base payment plus the average expected yield (5.5 tons per acre) multiplied by the incentive payment. Since the bidders were informed that this formula would be used to calculate total per-acre bids, the bids resembled a first-price sealed bid auction.

While the multiyear commitment needed from growers made written contracts a necessity, a conscious effort was made to avoid an overly burdensome or legalistic contract. Thus, all bidders executed a three-page written contract consisting of the University's standard form plus one page of additional terms (Epplin et al. 2007). The solicitation of bids began with an informational meeting that included: presentations on the project, switchgrass and switchgrass production (including a switchgrass production budget), the contract, and the bidding process. The forms needed for submitting a bid were distributed at the end of the meeting. County extension educators also took copies back to their counties for distribution to other interested individuals.

Bidders submitted a minimum and maximum acreage, with the University reserving the right to choose any acreage within that range. Due to the limited budget and the desire to contract with a number of different growers, the bidders were clearly informed that bids with relatively small acreages (i.e., approximately ten acres each) would be favored. While most of the bids proposed slightly more acreage than desired, few were for substantially more. However, one bidder was excluded because he offered a minimum of 70 acres. Anecdotal evidence also indicated that the low acreage allotments dissuaded some producers from bidding. In general, bidders seemed to understand the bidding process and the payment structure, and variation in the bids and per-acre amounts conformed to expectations.

[FIGURE 1 OMITTED]

Results--Land Lease Structure

Figure 1 illustrates the number of tons harvested per month for the eight-month and two-month harvest systems. Harvested tons differ across months because the number of harvest hours per day varies with average day length, and the number of harvest days varies with expected weather. If harvest is restricted to July and August, more than 390,000 tons would be scheduled for harvest in July and an additional 345,000 tons in August. If harvest could be spread over eight months, only 135,000 tons would be scheduled for harvest in July. Relatively few tons are harvested in October because of weather-related constraints on the number of harvest days. The expected October harvest is 40,000 tons. As reported in table 1, the optimal number of harvest units for raking-baling-stacking increases from 19 for the eight-month harvest system to 56 for the two-month harvest system. The average investment in harvest machines increases from $10.8 to $26.7 million as the length of the harvest season declines from eight to two months.

Table 1 includes a summary of results for both harvest systems. The number of harvested acres per year, which would be the same as the number of leased acres required to fulfill plant needs, is greater for the eight-month harvest system. This result is because the expected harvestable yield of switchgrass declines five percent per month from September through February. While the 2,000 tons per day biorefinery is expected to process 700,000 tons per year, excess tonnage must be harvested to offset expected storage losses. Thus, the harvested tons requirement is greater for the two-month harvest system.

Land rental and field costs per ton are similar for the two systems. The estimated cost to lease the land and produce, harvest, and store the switchgrass is $36.88 per ton for the eight-month harvest system and $52.75 per ton ($16 more) for the two-month harvest system (figure 2). The length of the harvest window matters, and this finding illustrates the potential economic value of a wide harvest window. It also suggests a potential economic problem for biorefineries designed to use crop residues as exclusive feedstock. For example, Nielsen (1995) estimates that the average harvest window for corn stover in the upper Midwest is 40 days. If corn stover were used as a single feedstock, a rather substantial investment would be required in harvest machines.

[FIGURE 2 OMITTED]

Results--Production Contract Structure

A summary of the eleven bids received by the University of Tennessee is provided in table 2. Contracts for a total of 92 acres were awarded to Bidders 2 through 6. On the basis of his acreage requirements, Bidder 1 was excluded from consideration, even though he was the low-cost bidder. Thus, contract awards were based on securing as many acres and as many different growers as possible, given the budget and structure of the bids.

While an expected average annual yield of 5.5 tons per acre was used for awarding bids for the four-year contracts, research trials conducted in the region suggest that an average annual yield of 7 tons per acre is more appropriate for an eleven-year period (the expected life of a switchgrass planting). Assuming a yield of 7 tons per acre, the weighted average of the accepted bids is $54.70 per ton. However, if only the lowest cost bids (Bidders 1 and 2 from table 2) had been accepted, the weighted average for the 92 acres would have been $35.99 per ton. Thus, the bidding process provides at least two possible cost levels: the weighted means of the low-cost bids ($35.99 per ton) and of the accepted bids ($54.70 per ton).

There are a number of reasons why these amounts might differ from those that might be obtained by a biorefinery. The informal acreage restriction imposed as part of the bidding process probably limited the number of bids and increased bid amounts by limiting the bidders' abilities to take advantage of size economies. Other reasons why the bids might not be representative of the costs of switchgrass to a biorefinery include: increased familiarity with switchgrass production costs and yields; structure of the contract (multiyear commitment at a fixed price); competition among potential buyers; and bias either in favor of or against participation in University research and/or contracting with the University. Therefore, the average bid price necessary to contract for the 100,000 to 125,000 acres necessary to provide feedstock for a 2,000 tons per day biorefinery may differ from that required to contract for 92 acres.

Discussion

The objective of the research was to determine the cost to produce switchgrass for both a land-lease alternative and a farmer-contract alternative, to determine if NREL's estimated delivery cost of $35 per dry ton is realistic, and to identify likely challenges to the development of switchgrass as a dedicated energy crop. The low-cost bids of $35.99 per ton and accepted bids $54.70 per ton obtained in the Tennessee project are comparable to the $36.88 and $52.75 per ton estimates obtained for the eight-month and two-month harvest systems with the Oklahoma model. When the $12 to $12.54 per ton transportation costs are included, the estimated cost to deliver feedstock ranges from $48 to $67 per dry ton. This finding is consistent with Petrolia's (2006) estimated cost to deliver corn stover to a 2,045 tons per day biorefinery in Minnesota of $54 per dry ton (excluding a payment to the farmer).

The $48 to $67 estimate is substantially more than the NREL goal of $35. At 90 gallons per ton, the feedstock component estimated cost is $0.53 to $0.74 per gallon of ethanol, as opposed to NREL's goal of $0.39. The exercise provides some insight as to what would be necessary to achieve NREL's goal of $35 per ton of delivered feedstock. It is unlikely that land, storage, and harvest costs (with existing harvest technology) could be much less than those estimated for the land-lease, eight-month system. Reductions in cost necessary to achieve the $35 goal would most likely require increases in switchgrass yields per acre.

The structure of a mature cellulosic feedstock production and delivery system may not resemble the atomistic system that we observe for U.S. grain, oilseed, and fiber production. If the low-cost feedstock is a perennial with a long-stand life and wide harvest window, such as switchgrass, market forces may drive the structure toward vertical integration. For a mature industry, feedstock production, harvest, and transportation may be centrally managed and coordinated. Established stands of healthy switchgrass are expected to require minimal attention and thrive for years. It is expected that an annual application of fertilizer and a single harvest would be sufficient, in which case switchgrass production for use as a dedicated energy crop could evolve to resemble a vertically integrated timber production and processing business.

A major limitation is that neither study considers the potential year-to-year variability in switchgrass yields. It may be optimal for a biorefinery to maintain a feedstock buffer. The model considers existing harvest technology that illustrates the consequences of a narrow harvest window on the cost to deliver cellulosic biomass. Restricting harvest from eight to two months per year increased the estimated delivered cost by 33% from $49 to $65 per ton. This finding illustrates the potential economic value of a wide harvest window. It also suggests a potential economic problem for biorefineries designed to use crop residues as exclusive feedstock. A conversion system that could use a variety of feedstocks including crop residues, wood wastes, and dedicated energy crops such as switchgrass, would have several potential advantages including a wide harvest window and a more variable rural landscape.

References

Aden, A., M. Ruth, K. Ibsen, J. Jechura, K. Neeves, J. Sheehan, B. Wallace, L. Montague, A. Slayton, and J. Lukas. 2002. "Lignocellulosic Biomass to Ethanol Process Design and Economics Utilizing Co-current Dilute Acid Prehydrolysis and Enzymatic Hydrolysis for Corn Stover." NREL/TP-510-32438, U.S. Department of Energy, National Renewable Energy Laboratory, Golden, CO.

Doye, D., and R. Sahs. 2007. "Oklahoma Cropland Rental Rates: 2006-07." Cooperative Extension Service Report CR-230, Oklahoma State University.

English, B.C., D.G. De La Torre Ugarte, K. Jensen, C. Hellwinckel, J. Menard, B. Wilson, R. Roberts, and M. Walsh. 2006. "25 % Renewable Energy for the United States by 2025: Agricultural and Economic Impacts." Department of Agricultural Economics, University of Tennessee.

Epplin, F.M., C.D. Clark, R.K. Roberts, and S. Hwang. 2007. "AJAE Appendix: Challenges to the Development of a Dedicated Energy Crop." Unpublished manuscript. Available at: http://agecon.lib.umn.edu/.

Fuentes, R.G., and C.M. Taliaferro. 2002. "Biomass Yield Stability of Switchgrass Cultivars." In J. Janick and A. Whipkey, eds. Trends in New Crops and New Uses. Alexandria VA: ASHS Press, pp. 276-82.

Graham, R.L., L.J. Allison, and D.A. Becker. i996. "ORECCL-Oak Ridge Crop County Level Database." Environmental Sciences Division, Bioenergy Feedstock Development Program, Oak Ridge National Laboratory, December. Available at: http://bioenergy. ornl.gov/papers/bioen96/graham2.html

Mapemba, L.D., F.M. Epplin, C.M. Taliaferro, and R.L. Huhnke. 2007. "Biorefinery Feedstock Production on Conservation Reserve Program Land." Review of Agricultural Economics 29:227-46.

McKendry, R 2002. "Energy Production from Biomass (part 2): Conversion Technologies." Bioresource Technology 83:47-54.

McLaughlin, S., J. Bouton, D. Bransby, B. Conger, W. Ocumpaugh, D. Parrish, C. Taliaferro, K. Vogel, and S. Wullschleger. 1999. "Developing Switchgrass as a Bioenergy Crop." In J. Janick, ed. Perspectives on New Crops and New Uses. Alexandria VA: ASHS Press, pp. 282-99.

Mosier, N., C. Wyman, B. Dale, R. Elander, Y.Y. Lee, M. Holtzapple, and M. Ladisch. 2005. "Features of Promising Technologies for Pretreatment of Lignocellulosic Biomass." Bioresource Technology 96:673-86.

Nielsen, R.L. 1995. "Questions Relative to Harvesting & Storing Corn Stover." Agronomy Department, AGRY-95-09, Purdue University, September.

Pacheco, M. 2006. "Invited Testimony for the U.S. Senate Committee on Energy and Natural Resources." Golden, CO: National Renewable Energy Laboratory, June 19.

Perlack, R.D., L.L. Wright, A.F. Turhollow, R.L. Graham, B.J. Stokes, and D.C. Erbach. 2005. Biomass as Feedstock for a Bioenergy and Bioproducts Industry: The Technical Feasibility of a Billion-Ton Annual Supply. Washington DC: U.S. Department of Agriculture, U.S. Department of Energy.

Petrolia, D.R. 2006. "The Economics of Harvesting and Transporting Corn Stover for Conversion to Fuel Ethanol: A Case Study for Minnesota." Department of Applied Economics, Staff Paper P06-12, University of Minnesota, August. Service, R.F. 2007. "Cellulosic Ethanol: Biofuel Researchers Prepare to Reap a New Harvest." Science 315:1488-91.

Tembo, G., EM. Epplin, and R.L. Huhnke. 2003. "Integrative Investment Appraisal of a Lignocellulosic Biomass-to-Ethanol Industry." Journal of Agricultural and Resource Economics 28:611-33.

Thorsell, S.R., F.M. Epplin, R.L. Huhnke, and C.M. Taliaferro. 2004. "Economics of a Coordinated Biorefinery Feedstock Harvest System: Lignocellulosic Biomass Harvest Cost." Biomass and Bioenergy 27:327-37.

Wyman, C.E. 1994. "Ethanol from Lignocellulosic Biomass: Technology, Economics, and Opportunities." Bioresource Technology 50:3-15.

Francis M. Epplin is Charles A. Breedlove Professor of Agricultural Economics at Oklahoma State University, Christopher D. Clark is Assistant Professor and Roland K. Roberts is Professor at the University of Tennessee, and Seonghuyk Hwang is a graduate research assistant at Oklahoma State University.

The authors acknowledge the assistance of personnel of the Biobased Products and Energy Center at Oklahoma State University. This article benefited from comments provided by B. Wade Brorsen and Jeffrey Vitale. Remaining errors are the responsibility of the authors. Research supported by the Oklahoma Agricultural Experiment Station, Project H-2574, by USDA-CSREES Special Research Grant award 2005-34447-15711, by the Tennessee Agricultural Experiment Station, Project TEN00256, and the U.S. Department of Energy through the Tennessee Switchgrass Project GO14219.

This article was presented in a principal paper session at the AAEA annual meeting (Portland, OR, July 2007). The articles in these sessions are not subjected to the journal's standard refereeing process. Table 1. Estimated Costs, Number of Harvest Machines, Average Investment in Harvest Machines, Acres and Tons Harvested to Provide a Flow of Switchgrass Feedstock to a 2,000 Dry Tons per Day Biorefinery for Both a Two- and Eight-Month Harvest Season

Model Results Category Eight-Month Two-Month

(a) (b) Costs

Land rent cost ($/ton) 10.77 9.74

Field cost ($/ton) (c) 9.23 8.35

Harvest cost ($/ton) 16.30 33.09

Field storage cost ($/ton) 0.58 1.57

Total cost other than

transportation ($/ton) 36.88 52.75

Transportation cost ($/ton) 12.00 12.54

Total cost of delivered

feedstock ($/ton) 48.88 65.29 Other results

Harvest units for mowing (number) (d) 47 136

Harvest units for

raking-baling-stacking (number) (e) 19 56

Average investment in harvest

machines ($,000) 10,777 26,726

Harvest acres 128,665 119,657

Total biomass harvested (tons) 716,635 736,741 (a) The eight-month harvest season extends from July through February. (b) The two-month harvest season includes only July and August. (c) Field cost includes amortized establishment, maintenance, and fertilizer costs. (d) A harvest unit for mowing includes one worker, one mower, and one tractor. (e) A harvest unit for raking-baling-stacking includes seven workers, three rakes, three balers, six tractors, and one transport slacker. Table 2. Tennessee Farmer Bids to Produce, Harvest, and Collect Switchgrass

Base

Minimum Maximum Bid Bidder (a) Acres Acres ($/acre) 1 70 100 $200.00 2 10 20 $250.00 3 8 15 $225.00 4 10 50 $200.00 5 12 30 $250.00 6 20 100 $255.05 7 10 50 $250.00 8 10 20 $255.34 9 16 16 $200.00 10 10 15 $62.00 11 10 20 $900.00

Total per

Incentive Acre Bid

Bid (5.5 t/a) (7 t/a) Bidder (a) ($/ton) 1 $7.50 $241.00 $253.00 2 $0.00 $250.00 $250.00 3 $20.00 $335.00 $365.00 4 $30.00 $365.00 $410.00 5 $25.00 $388.00 $425.00 6 $25.00 $393.00 $430.00 7 $30.00 $415.00 $460.00 8 $30.00 $420.00 $465.00 9 $50.00 $475.00 $550.00 10 $110.00 $667.00 $832.00 11 $30.00 $1,065.00 $1,110.00

Average Bid

per Ton (b) Acres Bidder (a) (5.5 t/a) (7 t/a) Awarded 1 $43.86 $36.07 0 2 $45.45 $35.71 15 3 $60.91 $52.14 15 4 $66.36 $58.57 30 5 $70.45 $60.71 12 6 $71.37 $61.44 20 7 $75.45 $65.71 0 8 $76.43 $66.48 0 9 $86.36 $78.57 0 10 $121.27 $118.86 0 11 $193.64 $158.57 0 (a) Farmers whose bids were accepted were contracted to seed switchgrass, fertilize, control weeds, harvest once per year, and collect harvested bales. Seed was provided and farmers were required to load but not transport the bales off the farm. (b) Weighted-average of the accepted bids is $63.69 assuming an actual yield of 5.5 tons per acre and $54.70 assuming an actual yield of seven tons per acre.