Relentless Service™
The National Energy Modeling System
The National Energy Modeling System (NEMS) was used to analyze the potential for cellulose-based ethanol production under various technological scenarios, assuming either a continuation of the Federal ethanol subsidy through 2020 or expiration of the subsidy in 2008. NEMS is a computer-based modeling system of U.S. energy markets for the midterm period of 1998 to 2020. NEMS projects the production, imports, conversion, consumption, and prices of energy, subject to assumptions on macroeconomic and financial factors, world energy markets, resource availability and costs, behavioral and technological choice criteria, cost and performance characteristics of energy technologies, and demographics. The Petroleum Market Model (PMM), a submodule of NEMS, is a linear programming representation of refining. It represents the pricing of petroleum products and crude oil, product import activity, and domestic refinery operations, subject to the demand for petroleum products, the prices of raw material inputs and imported petroleum products, the costs of investment, and the domestic production of crude oil and natural gas liquids.
The PMM includes an ethanol supply function that provides the linear program with supply curves for corn- and cellulose-based ethanol, allowing the PMM to project transportation ethanol supply throughout the NEMS forecast period. The ethanol model provides prices in the form of annual price-quantity curves. The curves, derived from an ethanol production cost function, represent the prices of ethanol at which associated quantities of transportation ethanol are expected to be available for production of E85 and ethyl tertiary butyl ether (ETBE), and for blending with gasoline.
The three PMM petroleum product supply regions are derived from the Petroleum Administration for Defense Districts (PADDs) as follows: region 1 represents PADD I, region 2 is an aggregate of PADDs II, III, and IV, and region 3 represents PADD V (Figure 5). The PMM demand regions are the nine U.S. Census divisions (Figure 6). Ethanol supply regions are also aggregated by Census division. The majority of ethanol supply derived from corn is located in Census divisions 3 and 4, with smaller amounts in divisions 6, 8, and 9, representing current supply. Cellulosic ethanol supplies become available in 2001 in Census divisions 2 and 7 at demonstration levels, and the majority of the projected growth (beginning in 2003) is in divisions 3, 4, and 9.
The largest growth in cellulose ethanol production is projected for Census divisions 3 and 4, the corn belt, where a tremendous supply of corn stover exists, as well as an established infrastructure for collecting and transporting agricultural materials. Census division 9 (mainly California) is projected to be the next largest producer of cellulose ethanol. It is assumed that ethanol will replace MTBE as the oxygenate for reformulated gasoline in California when the ban on MTBE takes effect in 2003, significantly increasing demand in the region. California's vast agricultural resources could sustain a cellulose ethanol industry of about 3 billion gallons per year.
The NEMS model was used to project potential biomass ethanol production in three different technological scenarios. The scenarios are based on the technologies described above and their associated cost savings potential. A reference case, similar to the Annual Energy Outlook 2000 (AEO2000) reference case, a high technology case, and a low technology case were examined. In addition, the effectiveness of the cost reductions projected by NREL was measured by the competitiveness of cellulose ethanol in the absence of the Federal subsidy.
The Federal Highway Bill of 1998 extended the current tax credit for ethanol through 2007 but stipulated reductions from the current 54 cents per gallon to 53 cents in 2001, 52 cents in 2003, and 51 cents in 2005. Although gasoline tax and tax credit provisions include "sunset" clauses that limit their duration, they have been extended historically. Therefore, a NEMS model assumption for AEO2000 was that the Federal subsidy would be extended at 51 cents per gallon through 2020.
State subsidies were also modeled in NEMS. While some ethanol-producing States do not subsidize ethanol, others offer tax incentives for gasoline blended with ethanol and for ethanol production, which vary from $0.10 to $0.40 per gallon (in nominal dollars). For modeling purposes, a volume-weighted average of $0.10 per gallon was used for corn-based ethanol in Census divisions 3 and 4.
The three technological simulations were run under two conditions to determine whether and at what price cellulose ethanol could remain competitive without the benefit of the Federal subsidy. Condition one extends the Federal subsidy at 51 cents per gallon through 2020, and condition two discontinues the subsidy in 2008.
The ethanol model uses a process costing approach to model the impacts of net feedstock production costs plus capital and operating costs associated with converting feedstock to ethanol. Corn feedstock prices were derived from USDA projections for the prices of corn and corn coproducts. Feedstock costs were calculated by subtracting the price of corn coproducts of wet and dry milling from the price of corn. Coproducts of wet milling were limited to corn gluten feed, corn gluten meal, and corn oil. Coproducts of dry milling consisted of distillers dried grains. USDA data were also used to vary corn and co-product prices as a function of ethanol demand. A study by Price et al. simulates the changes in production and consumption of major crops that would be caused by a change in corn ethanol production.
Cellulosic feedstock supply and prices are modeled in the NEMS Renewable Fuels Module. Biomass supply for ethanol competes with captive and noncaptive biomass markets. Captive markets pertain to users with dedicated biomass supplies who burn byproducts resulting from the manufacturing process. The noncaptive market includes the commercial, electric utility, transportation, and industrial sectors. The model calculates a supply schedule for each Census division, which defines the quantity and cost relationships of biomass resources accessible to all noncaptive consumers.
Biomass resources in the Renewable Fuels Module are an aggregation of forest products, wood wastes, crop residues, and energy crops. The forest products data were developed from U.S. Forest Service data, wood residue data were assembled from State and regional agency reports by Antares Group, Inc., and crop residue data were developed by Oak Ridge National Laboratory. Separate data for energy crops were compiled from an Oak Ridge National Laboratory database for each model year, 2010-2020, and added to the sum from the three other categories. The maximum share of cultivated cropland that would be used for energy crops was about 10 percent. A resource-related cost adjustment factor was also imposed to treat competing uses of the resource. For example, land could be used for other fiber or food crops, or the wood could be used for construction, at alternate prices. Figure 7 illustrates the composite U.S. total supply curve in 2010 for the first 50 million dry tons of biomass.
The methods of ethanol conversion assumed for this forecast varied across technological scenarios and were chosen according to their potential for cost reduction. Cumulative cost savings as a result of process improvements were based on NREL projections for each technology, calculated from a base conversion cost of $0.91 per gallon. Currently, there are several projects underway to produce ethanol from cellulose using either concentrated or dilute sulfuric acid hydrolysis technology. The low technology case assumed that the technology would continue to be used throughout the forecast period, and that process improvements would provide cost savings of 16 cents per gallon of ethanol by 2015. The countercurrent hydrolysis approach was chosen for the reference case technology. The countercurrent process improves on the dilute acid process, providing potential production cost savings of 30 cents per gallon of ethanol by 2015. The most advanced conversion process, with the greatest potential for cost reduction, is the enzymatic hydrolysis process. This process was assumed for the high technology case, with production cost savings of 60 cents per gallon of ethanol by 2015. Figure 8 compares ethanol price projections in the three technology cases with motor gasoline prices in the reference, low, and high world oil price cases.
Mansfield's assumptions in functional notation are given by:
[n(t+1) - n(t)]/[N - n(t)]=f(B, S, n(t)/N),
where
N=the total number of firms that may adopt the innovation,
n(t)=the total number of firms that have adopted the innovation by time t,
B=profitability of the innovation relative to other investments, and
S=the size of the investment needed to install the technology.
Mansfield then takes the first nonconstant term of the Taylor's expansion for f to rewrite the hypothesis as a differential equation:
dn(t)/dt=2 n(t)/N [N - n(t)] ,
where the constant, 2, consists of the terms
2=Z + a1 B + a2S.
Mansfield assumes, because of the limited number of innovations, that the coefficients of profitability and investment are constant over industries. The average payout period required by the firms to justify investments divided by the average payout period for the innovation is used as a measure of B. To measure S, he uses the average initial investment in the innovation as a percentage of the average total assets of the firms. Using these data, he obtains a least squares estimate of the parameters, resulting in the equation:
2=Z + 0.53 B - 0.027 S (r=0.997) ,
(0.015) (0.014) where the constants for the four industries (Z) are: -0.57 (coal mining), -0.52 (iron and steel), -0.59 (railroads), and -0.29 (brewing).
In his followup work, Blackman revised the model so that the extent of substitution was defined in terms of market share captured by the new technology rather than in terms of the cumulative number of firms employing the innovation. He applied the model to describe innovations dynamics in the commercial jet engine market and in the electrical utility and automotive sectors.
Blackman's market share formulation is given by:
N(t)=1 / [1 + exp(-k - 2t)] ,
where
N(t)=market share of new product, and
k=constant determined by initial conditions.
In the absence of historical data, Blackman suggested the use of an Innovation Index to estimate Z. The Innovation Index measures the relative propensity toward innovation in various industrial sectors of the U.S. economy. The index is derived from input variables that reflect the extent to which resources are allocated to achieve innovation in selected industrial sectors and output variables that measure the extent to which new product and process innovation is achieved. Blackman hypothesized that a relationship might exist between the value of Z for an industrial sector and the value of the Innovation Index for that sector. The hypothesis was tested using Z values from the steel, food and kindred products, aerospace, automotive, and electrical machinery sectors.
The following regression equation was obtained:
Z=0.2221 I - 0.3165 (r=0.92) ,
(0.0645)
where
I=the industry-specific Innovation Index.
Blackman computed the Innovation Index for 12 industrial sectors (Table 1). A positive Innovation Index indicates a strong tendency for an industry to innovate; a negative value indicates a weaker tendency for innovation.
Blackman's market share equation was used in NEMS to predict the rate of capacity expansion of cellulosic ethanol production. The cellulosic ethanol production capacity in year t is equal to the share of the market achieved in that year, N(t), times the total potential market for ethanol. The total ethanol market is defined as the sum of the potential gasohol market (10 percent blending of all traditional gasoline), the RFG oxygenate market, and the wintertime oxygenated gasoline market (approximately 12 billion gallons). The market penetration algorithm begins when the market share has reached 3 percent of the total market.
The constant k was determined from the initial condition; that is, at t0=0, the market share N(t0)=0.03. An initial growth rate of 12 percent per year (the approximate growth rate of corn-based ethanol production) was used to reach the 3-percent market penetration threshold. The parameters I, B, and S were assumed to vary across technological scenarios. The range for I was selected around the petroleum industry index (-0.64). The profitability index B increased from the low technology to the high technology case, reflecting the reduced costs of ethanol production. Profitability also varied across Census division, being highest in Census division 9. Several factors led to this decision. It was assumed that ethanol would be the oxygenate to replace MTBE in California RFG, creating a large increase in demand (over 550 million gallons in 2003) in the high-value RFG market. Census divisions 3 and 4 supply the Midwestern gasohol market, a lower value product, and the East Coast market, where the cost of transporting ethanol would further reduce profitability. The size of investment, S, is the relative size of the investment as a fraction of the total value of the firm. The low technology case uses a 50-percent fraction, implying high risk. The reference case uses 25 percent of the firm's value, and the high technology case uses 10 percent.
An alternative high technology case with capacity limited only by feedstock availability was also run, to determine the price at which blending of ethanol with conventional gasoline would occur without the benefit of a Federal subsidy. In this case, gasoline blending is projected to resume in 2010 in Census divisions 3 and 4 (Midwest), when the cost of cellulose ethanol drops to $0.82 per gallon. Ethanol begins to penetrate other markets in 2014, when costs fall to $0.68 per gallon.
Interestingly, the value of ethanol varies depending on how it is blended with gasoline. The marginal value of ethanol is higher in the projections when it is used as an oxygenate for RFG than when it is used as a volume extender. The projected marginal value of ethanol increases by $0.04 per gallon in Census division 3 and by $0.13 per gallon in Census division 9 when RFG blending begins in 2003 (Figure 11). Ethanol is also used as an oxygenate for wintertime fuels in areas that mandate the use of high oxygen (2.7 percent) fuels. Although ethanol is more expensive, it competes favorably with MTBE because it can provide the 2.7 percent oxygen requirement with only about 50 percent of the volume of MTBE. The NEMS model projects that ethanol will maintain its wintertime market share in high oxygen gasoline even in the absence of the Federal subsidy.
Ethanol costs could be reduced dramatically if efforts to produce ethanol from biomass are successful. Biomass feedstocks, including forest residue, agricultural residue, and energy crops, are abundant and relatively inexpensive, and they are expected to lower the cost of producing ethanol and provide stability to supply and price. In addition, the use of corn stover would lend continued support to the U.S. corn industry. Analysis of NREL technological goals for cellulose ethanol conversion suggests that ethanol could compete favorably with other gasoline additives without the benefit of a Federal subsidy if the goals were achieved. Enzymatic hydrolysis of cellulose appears to have the most potential for achieving the goals, but substantial reductions in the cost of producing cellulase enzymes and improvements in the fermentation of nonglucose sugars to ethanol still are needed.
The ban on MTBE in California could provide additional incentives for the development of cellulose-based ethanol. If ethanol were used to replace MTBE in Federal RFG, demand for ethanol in California would increase by more than 550 million gallons per year. California has vast biomass resources that could support the additional demand. In addition, the cost of transporting Midwest ethanol would allow cellulosic ethanol to compete favorably in the market. Ultimately, ethanol's future in RFG could depend on whether Congress eliminates the minimum oxygen requirement included in the CAAA90. Without the minimum oxygen requirement, refiners would have more flexibility to meet RFG specifications with blending alternatives, such as alkylates, depending on an individual refinery's configuration and market conditions. Ethanol would still be valuable as an octane booster, however, and could make up for some of the lost volume of MTBE.
Significant barriers to the success of cellulose-derived ethanol remain. For example, it may be difficult to create strains of genetically engineered yeast that are hardy enough to be used for ethanol production on a commercial scale. In addition, genetically modified organisms may have to be strictly contained. Other issues include the cost and mechanical difficulties associated with processing large amounts of wet solids. Proponents of biomass ethanol remain confident, however, that the process will succeed and low-cost ethanol will become a reality.
Source: EIA
The PMM includes an ethanol supply function that provides the linear program with supply curves for corn- and cellulose-based ethanol, allowing the PMM to project transportation ethanol supply throughout the NEMS forecast period. The ethanol model provides prices in the form of annual price-quantity curves. The curves, derived from an ethanol production cost function, represent the prices of ethanol at which associated quantities of transportation ethanol are expected to be available for production of E85 and ethyl tertiary butyl ether (ETBE), and for blending with gasoline.
The three PMM petroleum product supply regions are derived from the Petroleum Administration for Defense Districts (PADDs) as follows: region 1 represents PADD I, region 2 is an aggregate of PADDs II, III, and IV, and region 3 represents PADD V (Figure 5). The PMM demand regions are the nine U.S. Census divisions (Figure 6). Ethanol supply regions are also aggregated by Census division. The majority of ethanol supply derived from corn is located in Census divisions 3 and 4, with smaller amounts in divisions 6, 8, and 9, representing current supply. Cellulosic ethanol supplies become available in 2001 in Census divisions 2 and 7 at demonstration levels, and the majority of the projected growth (beginning in 2003) is in divisions 3, 4, and 9.
The largest growth in cellulose ethanol production is projected for Census divisions 3 and 4, the corn belt, where a tremendous supply of corn stover exists, as well as an established infrastructure for collecting and transporting agricultural materials. Census division 9 (mainly California) is projected to be the next largest producer of cellulose ethanol. It is assumed that ethanol will replace MTBE as the oxygenate for reformulated gasoline in California when the ban on MTBE takes effect in 2003, significantly increasing demand in the region. California's vast agricultural resources could sustain a cellulose ethanol industry of about 3 billion gallons per year.
The NEMS model was used to project potential biomass ethanol production in three different technological scenarios. The scenarios are based on the technologies described above and their associated cost savings potential. A reference case, similar to the Annual Energy Outlook 2000 (AEO2000) reference case, a high technology case, and a low technology case were examined. In addition, the effectiveness of the cost reductions projected by NREL was measured by the competitiveness of cellulose ethanol in the absence of the Federal subsidy.
The Federal Highway Bill of 1998 extended the current tax credit for ethanol through 2007 but stipulated reductions from the current 54 cents per gallon to 53 cents in 2001, 52 cents in 2003, and 51 cents in 2005. Although gasoline tax and tax credit provisions include "sunset" clauses that limit their duration, they have been extended historically. Therefore, a NEMS model assumption for AEO2000 was that the Federal subsidy would be extended at 51 cents per gallon through 2020.
State subsidies were also modeled in NEMS. While some ethanol-producing States do not subsidize ethanol, others offer tax incentives for gasoline blended with ethanol and for ethanol production, which vary from $0.10 to $0.40 per gallon (in nominal dollars). For modeling purposes, a volume-weighted average of $0.10 per gallon was used for corn-based ethanol in Census divisions 3 and 4.
The three technological simulations were run under two conditions to determine whether and at what price cellulose ethanol could remain competitive without the benefit of the Federal subsidy. Condition one extends the Federal subsidy at 51 cents per gallon through 2020, and condition two discontinues the subsidy in 2008.
Assumptions
Feedstock
The ethanol model uses a process costing approach to model the impacts of net feedstock production costs plus capital and operating costs associated with converting feedstock to ethanol. Corn feedstock prices were derived from USDA projections for the prices of corn and corn coproducts. Feedstock costs were calculated by subtracting the price of corn coproducts of wet and dry milling from the price of corn. Coproducts of wet milling were limited to corn gluten feed, corn gluten meal, and corn oil. Coproducts of dry milling consisted of distillers dried grains. USDA data were also used to vary corn and co-product prices as a function of ethanol demand. A study by Price et al. simulates the changes in production and consumption of major crops that would be caused by a change in corn ethanol production.
Cellulosic feedstock supply and prices are modeled in the NEMS Renewable Fuels Module. Biomass supply for ethanol competes with captive and noncaptive biomass markets. Captive markets pertain to users with dedicated biomass supplies who burn byproducts resulting from the manufacturing process. The noncaptive market includes the commercial, electric utility, transportation, and industrial sectors. The model calculates a supply schedule for each Census division, which defines the quantity and cost relationships of biomass resources accessible to all noncaptive consumers.
Biomass resources in the Renewable Fuels Module are an aggregation of forest products, wood wastes, crop residues, and energy crops. The forest products data were developed from U.S. Forest Service data, wood residue data were assembled from State and regional agency reports by Antares Group, Inc., and crop residue data were developed by Oak Ridge National Laboratory. Separate data for energy crops were compiled from an Oak Ridge National Laboratory database for each model year, 2010-2020, and added to the sum from the three other categories. The maximum share of cultivated cropland that would be used for energy crops was about 10 percent. A resource-related cost adjustment factor was also imposed to treat competing uses of the resource. For example, land could be used for other fiber or food crops, or the wood could be used for construction, at alternate prices. Figure 7 illustrates the composite U.S. total supply curve in 2010 for the first 50 million dry tons of biomass.
Technology/Costs
Conversion plant process costs (capital and operating) were assumed to be independent of production quantities. Plant size was considered in the overall cost of production, but it was assumed that savings from economies of scale would be offset by increased costs for feedstock collection. The operating costs (exclusive of energy) and capital costs for corn feedstocks were assumed to be constant over time. The amount of energy required to convert corn to ethanol, taken from Wang, was assumed to decrease linearly over time. Prices for coal and natural gas consumed during the conversion process were provided from the NEMS Coal Market Module and Natural Gas Transmission and Distribution Module, respectively. Total corn ethanol cost in the model was computed to be approximately $1.10 per gallon in 2000. The conversion and capital cost data for cellulose, derived from Wooley et al., were assumed to decrease over time at rates that varied across the technological scenarios. Wooley estimates production costs for a plant with a capacity of 2,000 tons per day (approximately 50 million gallons of ethanol) at $0.77 to $1.04 per gallon. An average of $0.91 per gallon was assumed as the initial cost for year 2000, resulting in a total cost for cellulosic ethanol production of approximately $1.29 per gallon. All costs are given in 1998 dollars.The methods of ethanol conversion assumed for this forecast varied across technological scenarios and were chosen according to their potential for cost reduction. Cumulative cost savings as a result of process improvements were based on NREL projections for each technology, calculated from a base conversion cost of $0.91 per gallon. Currently, there are several projects underway to produce ethanol from cellulose using either concentrated or dilute sulfuric acid hydrolysis technology. The low technology case assumed that the technology would continue to be used throughout the forecast period, and that process improvements would provide cost savings of 16 cents per gallon of ethanol by 2015. The countercurrent hydrolysis approach was chosen for the reference case technology. The countercurrent process improves on the dilute acid process, providing potential production cost savings of 30 cents per gallon of ethanol by 2015. The most advanced conversion process, with the greatest potential for cost reduction, is the enzymatic hydrolysis process. This process was assumed for the high technology case, with production cost savings of 60 cents per gallon of ethanol by 2015. Figure 8 compares ethanol price projections in the three technology cases with motor gasoline prices in the reference, low, and high world oil price cases.
Capacity
An important modeling consideration for the forecast of ethanol production from cellulose is the rate of capacity growth over the forecast period. Capacity expansion rates were projected using an algorithm derived from the Mansfield and Blackman statistical models of new technology market penetration. Mansfield investigated the factors that cause an innovation to spread through an industry. He examined the rate of substitution between time periods t and t+1 and hypothesized that the proportion of firms at time t that introduce the innovation by time t+1 is a function of: (1) the proportion of firms that have already introduced it at time t, (2) the profitability of the innovation relative to other investments, and (3) the size of the investment required to install the technology. He developed a deterministic model and fitted and tested it against data for 12 innovations in 4 industries.[n(t+1) - n(t)]/[N - n(t)]=f(B, S, n(t)/N),
where
N=the total number of firms that may adopt the innovation,
n(t)=the total number of firms that have adopted the innovation by time t,
B=profitability of the innovation relative to other investments, and
S=the size of the investment needed to install the technology.
Mansfield then takes the first nonconstant term of the Taylor's expansion for f to rewrite the hypothesis as a differential equation:
dn(t)/dt=2 n(t)/N [N - n(t)] ,
where the constant, 2, consists of the terms
2=Z + a1 B + a2S.
(0.015) (0.014)
In his followup work, Blackman revised the model so that the extent of substitution was defined in terms of market share captured by the new technology rather than in terms of the cumulative number of firms employing the innovation. He applied the model to describe innovations dynamics in the commercial jet engine market and in the electrical utility and automotive sectors.
N(t)=1 / [1 + exp(-k - 2t)] ,
where
N(t)=market share of new product, and
k=constant determined by initial conditions.
In the absence of historical data, Blackman suggested the use of an Innovation Index to estimate Z. The Innovation Index measures the relative propensity toward innovation in various industrial sectors of the U.S. economy. The index is derived from input variables that reflect the extent to which resources are allocated to achieve innovation in selected industrial sectors and output variables that measure the extent to which new product and process innovation is achieved. Blackman hypothesized that a relationship might exist between the value of Z for an industrial sector and the value of the Innovation Index for that sector. The hypothesis was tested using Z values from the steel, food and kindred products, aerospace, automotive, and electrical machinery sectors.
Z=0.2221 I - 0.3165 (r=0.92) ,
(0.0645)
where
I=the industry-specific Innovation Index.
Blackman computed the Innovation Index for 12 industrial sectors (Table 1). A positive Innovation Index indicates a strong tendency for an industry to innovate; a negative value indicates a weaker tendency for innovation.
Blackman's market share equation was used in NEMS to predict the rate of capacity expansion of cellulosic ethanol production. The cellulosic ethanol production capacity in year t is equal to the share of the market achieved in that year, N(t), times the total potential market for ethanol. The total ethanol market is defined as the sum of the potential gasohol market (10 percent blending of all traditional gasoline), the RFG oxygenate market, and the wintertime oxygenated gasoline market (approximately 12 billion gallons). The market penetration algorithm begins when the market share has reached 3 percent of the total market.
The constant k was determined from the initial condition; that is, at t0=0, the market share N(t0)=0.03. An initial growth rate of 12 percent per year (the approximate growth rate of corn-based ethanol production) was used to reach the 3-percent market penetration threshold. The parameters I, B, and S were assumed to vary across technological scenarios. The range for I was selected around the petroleum industry index (-0.64). The profitability index B increased from the low technology to the high technology case, reflecting the reduced costs of ethanol production. Profitability also varied across Census division, being highest in Census division 9. Several factors led to this decision. It was assumed that ethanol would be the oxygenate to replace MTBE in California RFG, creating a large increase in demand (over 550 million gallons in 2003) in the high-value RFG market. Census divisions 3 and 4 supply the Midwestern gasohol market, a lower value product, and the East Coast market, where the cost of transporting ethanol would further reduce profitability. The size of investment, S, is the relative size of the investment as a fraction of the total value of the firm. The low technology case uses a 50-percent fraction, implying high risk. The reference case uses 25 percent of the firm's value, and the high technology case uses 10 percent.
Results Federal Subsidy Extended to 2020
Benefitting from the assumed continuation of the Federal ethanol subsidy, gasoline blending of ethanol (in gasohol and RFG) is projected to increase by 1.4 percent per year from 2000 to 2020 in the reference case (Figure 9). Total U.S. cellulose ethanol production is projected to increase by 22 percent per year, reaching 850 million gallons by 2020 (Figure 10). Because cellulosic ethanol production capacity in Census division 9 does not grow sufficiently to meet California RFG demand, supplies of ethanol from the Midwest are needed to meet demand in Census division 9.Federal Subsidy Eliminated in 2008
When the Federal ethanol subsidy is assumed to be eliminated in 2008, gasohol and RFG blending with ethanol ceases in all three technology cases. Biomass ethanol production still is projected to grow in the reference and high technology cases, however, replacing the more expensive corn ethanol to meet California RFG and E85 demand. (A NEMS model assumption for this study was that demand for E85 would remain fixed and that RFG oxygenate demand in California would be met with ethanol.) In the low technology case, the projected growth of biomass ethanol production is similar under the subsidy extension and subsidy elimination assumptions, occurring only in Census division 9 to meet the required California RFG demand. Conventional gasoline blending of ethanol is projected to resume in the high technology case by 2018, when capacity begins to exceed the required demand for ethanol in RFG and E85.An alternative high technology case with capacity limited only by feedstock availability was also run, to determine the price at which blending of ethanol with conventional gasoline would occur without the benefit of a Federal subsidy. In this case, gasoline blending is projected to resume in 2010 in Census divisions 3 and 4 (Midwest), when the cost of cellulose ethanol drops to $0.82 per gallon. Ethanol begins to penetrate other markets in 2014, when costs fall to $0.68 per gallon.
Interestingly, the value of ethanol varies depending on how it is blended with gasoline. The marginal value of ethanol is higher in the projections when it is used as an oxygenate for RFG than when it is used as a volume extender. The projected marginal value of ethanol increases by $0.04 per gallon in Census division 3 and by $0.13 per gallon in Census division 9 when RFG blending begins in 2003 (Figure 11). Ethanol is also used as an oxygenate for wintertime fuels in areas that mandate the use of high oxygen (2.7 percent) fuels. Although ethanol is more expensive, it competes favorably with MTBE because it can provide the 2.7 percent oxygen requirement with only about 50 percent of the volume of MTBE. The NEMS model projects that ethanol will maintain its wintertime market share in high oxygen gasoline even in the absence of the Federal subsidy.
Conclusion
Ethanol has enjoyed some success as a renewable fuel, primarily as a gasoline volume extender and also as an oxygenate for high-oxygen fuels, an oxygenate in RFG in some markets, and potentially as a fuel in flexible-fuel vehicles. A large part of its success has been the Federal ethanol subsidy. With the subsidy due to expire in 2008, however, it is not clear whether ethanol will continue to receive political support. Thus, the future of ethanol may depend on whether it can compete with crude oil on its own merits.Ethanol costs could be reduced dramatically if efforts to produce ethanol from biomass are successful. Biomass feedstocks, including forest residue, agricultural residue, and energy crops, are abundant and relatively inexpensive, and they are expected to lower the cost of producing ethanol and provide stability to supply and price. In addition, the use of corn stover would lend continued support to the U.S. corn industry. Analysis of NREL technological goals for cellulose ethanol conversion suggests that ethanol could compete favorably with other gasoline additives without the benefit of a Federal subsidy if the goals were achieved. Enzymatic hydrolysis of cellulose appears to have the most potential for achieving the goals, but substantial reductions in the cost of producing cellulase enzymes and improvements in the fermentation of nonglucose sugars to ethanol still are needed.
The ban on MTBE in California could provide additional incentives for the development of cellulose-based ethanol. If ethanol were used to replace MTBE in Federal RFG, demand for ethanol in California would increase by more than 550 million gallons per year. California has vast biomass resources that could support the additional demand. In addition, the cost of transporting Midwest ethanol would allow cellulosic ethanol to compete favorably in the market. Ultimately, ethanol's future in RFG could depend on whether Congress eliminates the minimum oxygen requirement included in the CAAA90. Without the minimum oxygen requirement, refiners would have more flexibility to meet RFG specifications with blending alternatives, such as alkylates, depending on an individual refinery's configuration and market conditions. Ethanol would still be valuable as an octane booster, however, and could make up for some of the lost volume of MTBE.
Significant barriers to the success of cellulose-derived ethanol remain. For example, it may be difficult to create strains of genetically engineered yeast that are hardy enough to be used for ethanol production on a commercial scale. In addition, genetically modified organisms may have to be strictly contained. Other issues include the cost and mechanical difficulties associated with processing large amounts of wet solids. Proponents of biomass ethanol remain confident, however, that the process will succeed and low-cost ethanol will become a reality.
Source: EIA