Minimizing Land Use and Nitrogen Intensity of Bioenergy

Apr 26, 2010 - Solar energy conserved in biomass: Sustainable bioenergy use and reduction of land use change. Niclas S. Bentsen , Ian M. Møller. Rene...
61 downloads 4 Views 1MB Size
Download PDF
Environ. Sci. Technol. 2010, 44, 3932–3939

Minimizing Land Use and Nitrogen Intensity of Bioenergy SHELIE A. MILLER* Environmental Engineering and Earth Sciences, Clemson University 342 Computer Court, Anderson, South Carolina 29625

Received August 8, 2009. Revised manuscript received March 31, 2010. Accepted April 8, 2010.

The environmental impacts of bioenergy products have received a great deal of attention. Life cycle analysis (LCA) is a widely accepted method to quantify the environmental impacts of products. Conducting comprehensive LCAs for every possible bioenergy alternative is difficult because of the sheer magnitude of potential biomass sources and energy end products. The scopes of LCAs are often simplified to compare multiple products on the basis of greenhouse gas emissions and net energy balances, and may neglect equally important considerations such as nitrogen and land use. This study determines the most desirable energy crops on the basis of nitrogen and land use. The theoretical minimum nitrogen and land use requirements of fourteen bioenergy feedstocks are evaluated. These results can help prioritize certain feedstock crops for more in-depth life cycle analyses and can be used to inform policies on dedicated energy crops. The results of the study indicate that sugar cane has the best nitrogen and land use profile of the analyzed feedstocks. Sugar cane is the largest contributor to bioenergy production worldwide and is an effective policy choice from a nutrient and land use perspective. Conversely, soybeans and rapeseed are the least effective biomass sources with respect to land use and nitrogen requirements, yet these crops are also used to meet biofuel production targets worldwide. These results indicate current energy policies either do not consider or undervalue nitrogen and land use impacts, which could lead to unsustainable recommendations. Interestingly, when both nitrogen and land intensity are taken into account, reasonably small differences are seen between the remainder of the analyzed feedstocks, indicating an inherent trade-off between energy yield and nitrogen impacts.

Introduction A comprehensive assessment of bioenergy must take into account all major environmental issues of bioenergy production. Carbon emissions, nitrogen emissions, water use, and land use are issues of major concern for bioenergy technologies. Considerable attention has been given to carbon fluxes in bioenergy systems, including direct emissions throughout the life cycle and indirect emissions resulting from land use change (1-9). This trend is largely motivated by discussions of net energy balances and legislation requiring greenhouse gas savings throughout the life cycle (10-12). Significantly fewer studies have been conducted on the nitrogen, water, and land use aspects of bioenergy (13-19). The intention of this study is to provide * Corresponding author phone: 864-656-5572; fax: 864-656-0672; e-mail: [email protected]. 3932

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 10, 2010

baseline data that can be used in bioenergy LCA and act as a starting point to systematically evaluate preferable bioenergy crops. Bioenergy has many potential forms including combustion fuels (ethanol, biodiesel), electricity and heat generation, use in fuel cells, and gasification products (H2, butanol, etc.). Each bioenergy alternative can be generated from a number of biomass sources, resulting in different life cycle impacts. Much of the bioenergy discussion has focused on specific end products such as ethanol, without critically examining feedstock choices and the optimal use of those feedstocks. The corn-soybean rotation is predominant in U.S. agriculture, but it is unclear whether a system that is intended to maximize food production is the same as one that optimizes production of both food and fuel. Although ethanol production will continue to expand, there is not significant evidence to suggest that ethanol production from corn constitutes the most effective use of biomass resources in the United States. Although ethanol can be made from alternate feedstocks, such as switchgrass or woodchips, it has been suggested that electricity generation may be a more effective use of these resources from both economic and environmental perspectives (20). A systematic approach is needed to better inform policy decisions regarding which bioenergy feedstocks should be promoted and the best use of these resources. Compilation of comprehensive LCAs for all major impacts on all possible combinations of biomass feedstock and bioenergy end product is difficult because of the extensive data burden associated with such an effort. This method presents a basis on which potential biomass sources can be evaluated to determine the feedstocks with the fewest nitrogen and land requirements. This research focuses solely on biomass cultivation, assuming that creation of a sustainable biofuels industry relies upon sustainable feedstock selection. Once the initial impacts of biomass cultivation are better understood, more in depth LCAs can be conducted to determine the most effective use of these resources. This approach may be used to screen feedstock candidates with the least nitrogen and land use intensities; however, complete LCAs are still required to fully understand the environmental impacts of an individual bioenergy option. This study differs from LCA in that it calculates threshold values for entire biomass systems rather than calculating flows for individual products. In many cases, biomass systems can produce multiple coproducts. During the LCA process, the environmental impacts of the system are allocated to each coproduct. Allocation procedures used in LCA may obscure the true environmental costs of an entire system since the overall environmental impact is often divided over a number of coproducts (21). Allocation is avoided in this study for two reasons. First, the study focuses on the overall output of the entire system. The sum of all coproducts (fuel and food) will not exceed the maximum energy yield calculated in this paper. If the agricultural system is to be maximized for both food and fuel production, threshold energy yields need to be taken into account. Second, allocation can only be determined once the products of the system are identified. The coproducts may change for each bioenergy feedstock depending on the ultimate energy end point. For example, the corn profile will change depending on whether the residual corn stover is collected and used. Nitrogen and land use impacts are used as a screening tool since the majority of the impacts occur in a single life cycle stage of bioenergy products. Figure 1 shows generic life cycle results for bioenergy products (22). The height of the columns indicates impacts along the life cycle on a high/ 10.1021/es902405a

 2010 American Chemical Society

Published on Web 04/26/2010

FIGURE 1. Generic life cycle assessment of bioenergy products. low scale. The figure does not represent actual numerical values since inventories vary greatly among bioenergy products. Notably, nitrogen and land use impacts reside almost entirely within the agricultural production stage of the life cycle. Therefore, nitrogen and land use intensity can be generically defined for a variety of feedstocks irrespective of the ultimate end product because the only significant flows occur within the cultivation stage. Carbon and water fluxes, on the other hand, cannot be characterized in the same manner since there are significant fluxes in more than one life cycle stage, and are thus path dependent. Water can be an important consideration in both the cultivation and manufacturing stages of a bioenergy product (19). Water impacts are highly variable depending on the crop, location of the agricultural system, and the product produced. For example, ethanol production requires significant quantities of water, whereas producing biomass briquettes for combustion does not (19). An analysis of the entire life cycle must be conducted to determine the full carbon and water implications of any bioenergy option, and a feedstock screening tool cannot be employed the same way for water and carbon. The amount of land consumed by bioenergy production is a major issue. One of the key concerns is that bioenergy will compete with food supplies. In addition, growth of biofuels in one region of the world may cause a displacement of agriculture to sensitive ecosystems in other parts of the world (23, 24). Unlike fossil energy sources, biomass is very diffuse and requires a relatively large amount of land to grow energy. To have the least impact on global land use patterns, biofuel feedstocks with the highest energy yields per area should be chosen to minimize the land use intensity of biofuels. In addition to concerns regarding land use, agriculture is largely responsible for increased reactive nitrogen cycling. Human activities have disrupted the nitrogen cycle to a larger extent than the carbon cycle (25). Once in a reactive state, excess nitrogen cycles through its oxidation states and can participate in a variety of environmental impacts (26). Excess nitrogen in the environment contributes to human health impacts, smog formation, acid rain, eutrophication and coastal hypoxia, and climate change. Minimization of total nitrogen is desirable when the primary objective is production of a dedicated energy crop. In an energy system, nitrogen emissions are an unintentional byproduct and should be avoided. This is contrary to the traditional food production model, which values nitrogen for its nutritional properties. Both production of synthetic fertilizer and increased biological nitrogen fixation (BNF) from intensive cultivation of legumes are responsible for disruptions in the nitrogen cycle (27, 28). Cultivation of crops that use BNF is very advantageous for improved nitrogen management in crop rotations and often reduces overall nitrogen burdens. BNF has lower energy consumption

throughout the life cycle and better nitrogen uptake efficiencies within the soil when compared to synthetic fertilizer. Production of synthetic fertilizer is highly energy intensive, representing 1% of global energy consumption (29). These advantages of BNF are captured in the full-scale LCA process and are not included in this study. Of the feedstocks analyzed, soybeans obtain the most significant fraction of their nitrogen from BNF. Although BNF is generally seen as a preferable way for systems to obtain nitrogen, the effects of excess nitrogen within a system are the same regardless of origin. Nitrate runoff from feedlot operations causes the same water quality impacts whether the protein comes from synthetically fertilized corn or biologically fixed soybeans. Although some feedstocks have the ability to supply food and the supplied nitrogen may serve as nutrition up the food chain, the purpose of this study is to identify optimal bioenergy feedstocks. Therefore, all nitrogen is treated equally, regardless of origin and eventual use. This analysis calculates the minimum nitrogen and land required to produce 1000 GJ of unprocessed energy. These values identify the theoretical limits of bioenergy systems, presenting a best case scenario. It will never be possible to use less land or supply less nitrogen because the system is constrained by these values. Values for actual nitrogen and land requirements for any product will be significantly greater due to thermodynamic losses throughout the life cycle and nitrogen applied in excess to compensate for inefficiencies in nitrogen uptake. Because this study focuses on the theoretical minimums of nitrogen and land intensity, it differs from life cycle studies measuring actual inventory data for a specific product (18, 30). It cannot be assumed that conversions to the final product will be equivalent for all options. The additional land and nitrogen required for an end product is highly variable and requires a full LCA to calculate.

Methods The theoretical limits for potential energy generation, land use, and amount of nitrogen required by the system are calculated for fourteen feedstocks. The feedstocks compared in this study are either currently used for bioenergy or are candidate bioenergy feedstocks. They include: algae, corn, grain sorghum, oil palm, perennial grasses (miscanthus, switchgrass), rapeseed, short rotation trees (birch, poplar, willow), soybean, sugar beet, sugar cane, and sweet sorghum. All are agricultural or forestry products except for algae, which is grown in industrial operations. Two varieties of algae are shown at high lipid (70% by dry mass) and low lipid (30% by dry mass) concentrations. The two varieties of sorghum include currently cultivated varieties (grain sorghum) and modified varieties that maximize biomass with high sugar contents (sweet sorghum). These fourteen products are chosen to represent a diversity of materials over a variety of regions. Only a subset of the analyzed feedstocks can be grown in any given geographic region. The data presented for each feedstock is representative over its geographically appropriate range. Table 1 classifies the analyzed biomass feedstocks into categories with similar characteristics. Only the harvestable portion of biomass is considered in this analysis. The harvestable portion refers to the amount of the plant material that can be reasonably collected from a field or industrial operation and processed into an energy source. For soybean and rapeseed, the harvestable portion is the oil seeds since residue collection is infeasible. The harvestable portion of the oil palm is the fruit, which includes the mesocarp, endocarp, and kernel. The harvestable portion for corn, forestry products, switchgrass, miscanthus, sugar cane, grain sorghum, and sweet sorghum refers to the total VOL. 44, NO. 10, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3933

TABLE 1. Biomass Feedstocks Separated by Category category

included feedstocks

description

food crops cellulosic

corn, oil palm, rapeseed, soybean, grain sorghum species historically cultivated as sources of food or feed short rotation trees (birch, poplar, willow) species targeted as next-generation biofuels that employ a perennial grasses (miscanthus, switchgrass) more efficient (C4) photosynthetic pathway than many annual crops, which allows for greater yields and lower nitrogen requirements sugar crops sugar cane, sugar beets, sweet sorghum high-yielding species characterized by high sugar and low protein contents industrial algae organisms produced in industrial systems and characterized by high protein content and biomass yields; highly variable yields due to specific industrial conditions and algae strain

aboveground biomass. These products are either harvested in their entirety or have significant residual material that can be used in bioenergy production. Data for algae and sugar beet refer to total biomass yield. For algae, both the lipid fraction and cellular material have energetic value. Sugar beets have both above and below ground biomass in significant quantities that have potential energetic value. Depending on the feedstock, energy obtained can be used as food, feed, fuel, or some combination of the three, but no differentiation is made in this analysis. The maximum energy yield (MEY) is the maximum amount of raw energy that can be obtained from a unit of land. Energy yields are calculated by multiplying the harvestable yield by the high heating value (HHV) of each feedstock. This calculation gives the maximum energy that can be obtained from a biomass source. For combustion processes, the low heating value (LHV) is often chosen to take into account the vaporization of water during combustion; however, bioenergy options such as fuel cells and gasification do not rely on combustion, so HHV is used to calculate maximum possible energy. The MEY does not take into account energy losses during processing or energy used to create the biofuel. Actual net energy yields will be much lower because of thermodynamic losses and the energy used to grow, harvest, and process the biomass throughout the life cycle. The actual amount of net energy obtained from each feedstock will vary according to these parameters and can be calculated by the net energy balance method that has been extensively reported in the literature (5-9). Land use intensity is calculated by determining the amount of land required to supply a given amount of energy. It is the inverse of the calculated energy yield. For easier comparison, land intensity is calculated as the land needed to produce 1000 GJ of raw energy. The minimum nitrogen requirement (MNR) is the amount of reactive nitrogen that is removed with the harvestable yield. The MNR is calculated by multiplying the harvestable yield and the percent nitrogen composition. Nitrogen in nonharvestable plant residues left on the field are not included in the calculation. These are assumed to be reincorporated into the soil and recycled during the next growing period. In order to meet the MNR for an organism, nitrogen is often supplied by applying fertilizer in excess. Most plants are inefficient at extracting reactive nitrogen from soil, so the excess is directly released to the environment or incorporated into the soil organic matter. All reported values are annualized. Forestry products are normalized over the period from planting to harvest. Since crop yields are highly variable according to geography, climate, and agricultural practices, a range of values are presented for the land use and nitrogen intensities. The mean values are used for discussion purposes, but the reported range encompasses a representative variety of geographic regions and management practices. Calculations for each 3934

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 10, 2010

feedstock, baseline assumptions, and source information are contained in the Supporting Information. Ranking Procedure. Land use and nitrogen impacts are incommensurable and not easily converted into a single score (31). A standard Grid Analysis Method is used to determine a weighted score with respect to both categories. Grid Analysis is a standard tool of Multi-Criteria Decision Analysis (32). Each data point is divided by the overall column sum to determine its ranking for land use and nitrogen intensity. In cases where values are within a standard deviation, the feedstocks are given the same ranks. Resulting scores for land use and nitrogen intensity are averaged for each feedstock to give the weighted score. The weighted scores and ranks only have meaning in the context of this analysis and will change depending on the feedstocks compared. Using this method, equal weight is given to land use and nitrogen intensity. This is reasonable method for the purposes of the analysis since it allows the feedstocks to be compared directly to one another.

Results and Discussion Minimum Land Requirements. Figure 2 shows the land intensity of the biomass resources, indicating the amount of land needed to harvest 1000 GJ of raw energy. These data are calculated using the maximum energy yield over a range of growing conditions and regions. The land intensity values in Figure 2 can be roughly separated into the categories in Table 1. Land intensity is lowest for industrial algae and sugar crops. Sugar crops have high yields and the sugar content contributes to marginally higher heating values. Cellulosic feedstocks are quite variable depending on the yield, stand maturity at harvest, and planting density but are less land intensive than the food crop category in general. Oil palm, while categorized as a food crop, has a lower land intensity than the rest of the food or cellulosic feedstocks. Corn is the next best food crop, with a land intensity similar to many of the cellulosic species. Rapeseed, soybean, and grain sorghum have the lowest energy yields and are therefore the least desirable from a land intensity standpoint. Minimum Nitrogen Requirements. Figure 3 shows the minimum reactive nitrogen (MNR) necessary to support a range of energy yields. This nitrogen may be supplied via synthetic fertilizer applications or biological nitrogen fixation. The slopes of lines in Figure 3 are given as GJ/kg N, indicating the marginal energy yield per unit of nitrogen supplied. Steep slopes require only small quantities of additional nitrogen to obtain greater yields. Flatter slopes are less desirable since a greater quantity of nitrogen must be supplied to increase energy yield. An ideal bioenergy feedstock would have a steep slope and be located in the upper left corner. It can be clearly seen that sugar cane has reasonably small nitrogen requirements as well as a high energy yield. The remainder of the feedstocks show definite trade-offs between nitrogen re-

FIGURE 2. Hectares of land needed to produce 1000 GJ of raw energy.

FIGURE 3. Minimum nitrogen requirement (MNR) to generate a range of energy yields. Food crops indicated by solid black line, black fill; cellulosics have solid blue line, no fill; sugar crops have dotted black lines, red fill; algae has solid green line, green fill. Note the break in scale. quirements and the amount of energy they can generate. These data represent the best possible scenario and are the system limits for each feedstock. In practice, the slopes of the lines will be flatter and all the lines will migrate down and to the right when overall life cycle data are taken into account. The results in Figure 3 are also roughly grouped into the categories in Table 1. Algae has a biomass producing potential

orders of magnitude greater than the other biomass sources. It also requires the largest amount of reactive nitrogen. As indicated by their relatively flat slopes, food crops are nitrogen intensive and require significant additional nitrogen inputs for small increases in energy yield. As a whole, cellulosics have lower MNRs and greater energy yields than food crops. Sugar crops have both steep slopes and high energy yield potential, containing a large amount of simple sugars that VOL. 44, NO. 10, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3935

TABLE 2. Biomass Ranked on the Basis of Land Use and Nitrogen Intensity to Produce 1000 GJ Energya land use intensity nitrogen intensity weighted (ha/1000 GJ) rank (kg N/1000 GJ) rank score sugar cane willow miscanthus sugar beet oil palm birch poplar switchgrass corn sweet sorghum algae grain sorghum rapeseed soybean a

2.3 5.3 4.2 1.9 3.0 6.8 7.2 6.5 4.9 6.1

2 6 5 2 4 8 8 8 6 8

110 90 210 460 440 160 160 300 490 390

2 1 5 8 8 3 3 6 8 7

0.02 0.03 0.03 0.03 0.04 0.04 0.04 0.05 0.05 0.05

0.3 16.2

1 12

1100 1000

11 11

0.05 0.13

16.5 20.2

12 13

1400 3900

12 13

0.15 0.30

Lower numbers are preferable for weighted score.

can easily be converted to energy. Although sugar crops contain very small amounts of protein, they possess relatively high MNRs because of high yields. In some cases, the energy yield shown in Figure 3 are higher than values reported elsewhere (18, 33). These differences result from both the inclusion of harvestable residues and calculation of the maximum energy yield. Other studies may report net energy calculations or use different harvestable fractions. Because crops are not 100% efficient at taking up nitrogen, nitrogen is often supplied in excess to maximize yield. The nitrogen actually supplied to the system will be greater than that shown in Figure 3 and will depend on a variety of factors. The dynamics of nitrogen fertilization are complicated and not easy to generalize. Nitrogen uptake efficiency is dependent on crop type and changes with the rate of fertilization. Fertilization rates for each crop vary greatly by region and agricultural management practices. If nitrogen supply is insufficient, crop yield will be negatively impacted. For the data reported in this study, it is assumed that the crops receive adequate nitrogen supply to maintain the average reported yield. Because they are not nutrient limited, it is assumed they are obtaining their maximum yield potential. Sugar crops are excellent nitrogen scavengers and often do not receive high fertilization (34). There is some evidence that certain varieties or sugar cane or sugar beet have cooperative relationships with nitrogen fixing bacteria, which reduces overall synthetic nitrogen requirements (35, 36). Potential exists to promote cultivars that form symbiotic relationships with nitrogen fixing bacteria, thereby greatly reducing or eliminating the need for fertilization of sugar crops (35, 36). Because soybeans are known to fix nitrogen, the reported MNR for soybean is similar to what is expected to be actually obtained by the system. The reported forestry products also have a small differential between the MNR and the amount of applied synthetic nitrogen. Trees are excellent nitrogen scavengers, and fertilization of managed forests has only recently become a widespread practice (37). Nitrogen requirements are generally quite low and are acquired over a period of years. Food crops have the largest difference between supplied fertilizer and MNR, which causes excess nitrogen to be released to the environment (38). Feedstock Ranking. Table 2 summarizes the results of the analysis, showing average land use intensity and nitrogen intensity on the basis of producing 1000 GJ of raw energy. The weighted score determines the relative environmental performance of each feedstock on the basis of land use and nitrogen intensity. These scores are based on average data and can be used to understand general trends in the data. These scores can be used to compare nitrogen and land 3936

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 10, 2010

profiles of these fourteen feedstocks relative to each other. Absolute ranking of individual feedstocks should be avoided due to variability of nitrogen and land profiles and the sensitivity of these data to baseline assumptions. Because the weighted scores are calculated via a normalization procedure, these scores will obviously change according to the number and variety of feedstocks analyzed. Table 2 shows a distinct break between the first ten feedstocks and grain sorghum, rapeseed, and soybean. These three species are consistently ranked the lowest for both land and nitrogen intensity. Comparatively small differences are observed among the ranks of the remaining eleven feedstocks. These can be grouped according to their ranked values. Sugar cane is consistently ranked the best, followed by three groups containing three to four feedstocks each. Sensitivity Analysis. Nitrogen and land use intensity calculations are sensitive to assumptions of HHV, nitrogen content, moisture content, and harvestable yield. Each of these factors has a range or reasonable values depending on geographic region and management practices. A sensitivity analysis shows that the eleven most highly ranked species may change places with adjacent species when the assumptions are varied across their range; however, overall trends remain unchanged. For example, for the assumed range of data, sugar cane is always the most highly ranked. Soybean, rapeseed, and grain sorghum are always ranked the most poorly, regardless of assumptions. The short rotation trees, sugar beet, miscanthus, and oil palm are consistently among the highest-ranked species. Switchgrass, corn, and sweet sorghum also fall within a similar range and tend to be below the short rotation trees throughout their sensitivity range. These trends show that feedstocks can be prioritized, but specific geographic and management information is needed to narrow the ranges to make definitive comparisons. Because the variables are linearly related, the system is most sensitive to parameters that have ranges with the highest percent difference from the average. For example, nitrogen intensity calculations are highly sensitive to variations in nitrogen content. Sugar crops are especially sensitive to this parameter because the nitrogen content is a fraction of a percent and small variations change overall MNR significantly. Land intensity calculations are less sensitive than nitrogen intensity since energy content and yield ranges possess inherently less relative variability. Trade-off Analysis. There is an observed trade-off between land use and nitrogen intensity, which can be observed in Figure 4. Figure 4 only shows the top eleven feedstocks to better illustrate differences in data. A feedstock that has low land requirements tends to have reasonably high nitrogen requirements, and vice versa. This trade-off leads to the reasonably small differences in weighted scores. This tradeoff is particularly evident for algae. Algae has the lowest land footprint yet has one of the highest nitrogen requirements. It is important to recognize this trade-off in environmental assessments. When energy yield is used as a single criterion, major differences can be seen between the feedstocks that disappear when nitrogen is included in the consideration. From the perspective of nitrogen cycle disruption, dedicated soybeans and rapeseed are especially unfavorable for bioenergy production because of low yields and high reactive nitrogen requirements, even though the majority of soybean nitrogen is obtained via BNF. Interestingly, biodiesel produced from soybean and rapeseed tends to perform relatively well in life cycle assessment (13, 39). The life cycle assessment procedure allocates environmental impacts among coproducts. For oil seed crops, the majority of the environmental impact is allocated to the meal. While this is a reasonable allocation procedure consistent with LCA methods, it is important to highlight that, overall, growing the soybean and rapeseed feedstocks are fundamentally more intensive

FIGURE 4. Land use and nitrogen intensity for the analyzed feedstocks. than other crops due to much lower yields and higher nitrogen requirements. These intensities are justified when the crops are used primarily as a food source to capitalize on their high protein contents; however, the high nitrogen intensity is not favorable if production of these crops is primarily driven by policies aimed at expanding bioenergy feedstocks. Of course, it should be noted that this study does not factor in coproducts and compares the feedstock as a whole. As noted above, only a fraction of these feedstocks will be converted into bioenergy, with the meal portion used as a valuable food or feed resource. A full scale LCA accounts for the provision of multiple products and will allocate environmental impact accordingly, whereas this screening method assesses the feedstock as a whole without any form of coproduct allocation. A high protein content makes soybeans desirable from a food and feed standpoint, and soybeans can be part of an effective nitrogen management strategy when rotated with nitrogen intensive crops such as corn. Rapeseed, which is promoted in the European Union, may be an even less desirable crop since it obtains its protein via synthetic fertilizer rather than biological fixation, creating greater environmental impacts up the supply chain. While certainly not ideal, corn is the most preferable of the traditional row crops from both a nitrogen intensity and land use perspective, presumably because of its classification as a grass and its more efficient photosynthetic pathway. Oil derived from food production as a waste or byproduct has definite appeal for bioenergy production; however, this analysis concludes that soybean and rapeseed are not logical choices for crops dedicated and promoted specifically for bioenergy production. Algae as an Industrialized Biomass Source. Algae is grown in contained, industrialized systems, which makes it unique among bioenergy options. As an industrialized product, management of algae systems can be quite different than management of other bioenergy options. Depending on the system design, algae systems may be less susceptible to external variables, such as weather. Nonpoint runoff emissions are also not an issue. Algal systems are also not as geographically constrained as agricultural systems and can be implemented in areas that are not agriculturally productive. The unique aspects of algae systems require further discussion with respect to the issue of nitrogen management. Algae has nitrogen requirements an order of magnitude greater than the other feedstocks included in this study. The nitrogen requirements are not only because of the highyielding nature of the organism but also a low carbon to nitrogen ratio that approaches the Redfield ratio (40). The high nitrogen requirement can either have positive or negative impacts on the nitrogen cycle depending upon how algae systems are developed and managed. Nitrogen can be

recycled within the system and may be obtained from waste sources from municipal or animal operations (41). If the required nitrogen is supplied by a waste source, there may be a net decrease in reactive nitrogen emissions by using waste materials as inputs. In addition, algae from which the oil has been extracted may act as a high-protein animal feed source that could displace crops such as soybean, which has high land requirements. Conversely, addition of synthetic nitrogen will have unfavorable impacts. Although the high nitrogen requirements of algae are cause for concern, appropriate development and management can lead to positive impacts on the nitrogen cycle. Algal systems are still under development, and it is unclear exactly how these systems will be managed or where they will be implemented. Because of the uncertainty regarding these aspects, algae is treated the same as the other feedstock categories when ranking the results. Algae has the largest trade-off between land area and nitrogen requirements. The ranking of algae would change significantly if it could be shown that the nitrogen management would net overall system benefits. Land Quantity versus Land Quality. The environmental impacts of bioenergy are complex, and there are numerous other considerations not included in this analysis. One of the major issues is considerations of land quality. It is important to note that issues concerning land use extend beyond a simple calculation of area occupied by biomass. In addition to land quantity, the improvement or degradation of land resources is a major consideration with respect to bioenergy. Increased bioenergy production may lead to significant land use change globally, which in turn may impact elemental fluxes to and from the soil, erosion effects, and provision of ecosystem services. The effects of land use change will be specific to the biomass under cultivation and the prior use of land. For example, high-intensity agriculture that is converted to low-input mixed perennial grasses would result in an improvement of land quality (42), whereas conversion of native forest ecosystems to perennial grasses would degrade quality (43). Furthermore, industrialized systems, such as algae, are not constrained to arable lands like other biomass sources. Algae systems can be developed in areas that do not compete with agriculture, such as the desert or in the ocean. Alternatively, some perennial grasses are drought tolerant and can grow well on marginal lands not suitable for other agricultural purposes. These considerations cannot be captured by simple quantification of land area and are not reflected in the results of this study. Incorporation of land quality and ecosystem service metrics into LCA is an active research area that requires lengthy debate. Proposed methods to include ecosystem services and prior land occupation into LCA considerations are under development VOL. 44, NO. 10, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3937

(44). This study represents a first step to systematically include land use considerations, although calculation of ecosystem services is outside the scope of this paper. Inclusion in LCA and Policy Considerations. This study determines that both land use and nitrogen intensity must be considered when selecting bioenergy feedstocks and may have significant policy implications. For example, sugar beet grows in a fairly broad geographic range that overlaps with corn agriculture. Even though sugar beet has a significantly better profile with respect to land and nitrogen requirements, corn is the crop promoted by U.S. bioenergy policy (12). There are numerous factors that go into agricultural production decisions, although the major barrier to sugar beet intensification is not necessarily economic. It has been suggested that increased sugar beet production is impeded by a lack of infrastructure and policy incentives promoting other crops, such as corn (45). It must be emphasized that the method shown cannot replace LCA results but can be useful in better understanding overall land and nitrogen requirements for different biomass systems to determine theoretical limits. Nitrogen and land use intensity are certainly not the only considerations when taking into account the environmental implications of energy crops. Overall life cycle considerations must be included, such as land use change, net carbon emissions, impact on biodiversity, emissions of other nutrients (P, K), soil erosion, water and pesticide use, and impact of production and processing activities.

(11) (12)

(13)

(14)

(15)

(16)

(17)

(18)

(19)

(20)

Acknowledgments

(21)

Support for this research is provided by the National Science Foundation CAREER program #0845728.

(22)

Supporting Information Available Baseline assumptions, source information, and greater explanation of maximum energy yield and minimum nitrogen requirement calculations.This material is available free of charge via the Internet at http://pubs.acs.org.

Literature Cited (1) Shapouri, H.; Duffield, J. A.; Wang, M. Q. The Energy Balance of Corn Ethanol: An Update; United States Department of Agriculture, Economic Research Service: Washington, DC, 2002. (2) Borjesson, P. I. I. Energy analysis of biomass production and transportation. Biomass Bioenergy 1996, 11 (4), 305–318. (3) Kim, S.; Dale, B. E. Environmental aspects of ethanol derived from no-tilled corn grain: nonrenewable energy consumption and greenhouse gas emissions. Biomass Bioenergy 2005, 28 (5), 475–489. (4) Blottnitz, H. v.; Curran, M. A. A review of assessments conducted on bio-ethanol as a transportation fuel from a net energy, greenhouse gas, and environmental life cycle perspective. J. Cleaner Prod. 2007, 15, 607–619. (5) Lynd, L. R. Overview and evaluation of fuel ethanol production from cellulosic biomass: technology, economics, the environment, and policy. Annu. Rev. Energy Environ. 1996, 21 (1), 403– 465. (6) Kim, S.; Dale, B. E. Allocation procedure in ethanol production system from corn grain. Int. J. Life Cycle Assess. 2002, 7 (4), 237–243. (7) Shapouri, H.; Duffield, J. A.; Graboski, M. Estimating the Net Energy Balance of Corn Ethanol; Agricultural Economic Report 721; US Department of Agriculture, Economic Research Service: Washington, DC, 1995. (8) von Blottnitz, H.; Curran, M. A. A review of assessments conducted on bio-ethanol as a transportation fuel from a net energy, greenhouse gas, and environmental life cycle perspective. J. Cleaner Prod. 2007, 15, 607–619. (9) Hulsbergen, K. J.; Feil, B.; Biermann, S.; Rathke, G. W.; Kalk, W. D.; Diepenbrock, W. A method of energy balancing in crop production and its application in a long-term fertilizer trial. Agric., Ecosyst. Environ. 2001, 86, 303–321. (10) U.S. EPA Regulation of Fuels and Fuel Additives: Changes to Renewable Fuel Standard Program, 99th ed.; U.S. EPA: Wash3938

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 10, 2010

(23) (24)

(25)

(26)

(27) (28)

(29) (30)

(31) (32) (33)

(34)

(35)

(36)

ington, DC, 2009; Vol. 74;Federal Register 2009, 40 (CFR Part 80), 24904-25143. Anex, R.; Lifset, R. Assessing corn ethanol. J. Ind. Ecology 2009, 13 (4), 479–482. Energy independence and security act. Public Law 110-140, 110th Congres 2007; http://frwebgate.access.gpo.gov/cgi-bin/ getdoc.cgi?dbname)110_cong_public_laws&docid)f: publ140.110.pdf. Miller, S. A.; Landis, A. E.; Theis, T. L. Environmental tradeoffs of bio-based production. Environ. Sci. Technol. 2007, 41 (15), 5176–5182. Landis, A. E.; Miller, S. A.; Theis, T. L. Life cycle of the cornsoybean agroecosystem for biobased production. Environ. Sci. Technol. 2007, 41 (4), 1457–1464. van den Broek, R.; Meeusen, M.; Treffers, D.-J.; van Wijk, A.; Nieuwlaar, E.; Turkenburg, W. Green energy or organic food? A life-cycle assessment comparing two uses of set-aside land. J. Ind. Ecol. 2001, 5 (3), 65–87. Owens, J. W. Water resources in life-cycle impact assessment: Considerations in choosing category indicators. J. Ind. Ecol. 2001, 5 (2), 37–54. Renouf, M. A.; Wegener, M. K.; Nielsen, L. K. An environmental life cycle assessment comparing Australian sugarcane with US corn and UK sugar beet as producers of sugars for fermentation. Biomass Bioenergy 2008, 32 (12), 1144–1155. Dominguez-Faus, R.; Powers, S. E.; Burken, J. G.; Alvarez, P. J. The water footprint of biofuels: A drink or drive issue. Environ. Sci. Technol. 2009, 43 (9), 3005–3010. Berndes, G. Bioenergy and watersThe implications of largescale bioenergy production for water use and supply. Global Environ. Change 2002, 12 (4), 253–271. Campbell, J. E.; Lobell, D. B.; Field, C. B., Greater Transportation Energy and Ghg Offsets from Bioelectricity than Ethanol. Science Express May 7, 2009. Weidema, B. Avoiding co-product allocation in life-cycle assessment. J. Ind. Ecol. 2000, 4 (3), 11–33. Zah, R.; Faist, M.; Reinhard, J.; Birchmeier, D. Standardized and simplified life-cycle assessment (LCA) as a driver for more sustainable biofuels. Journal of Cleaner Production 2009, 17 (Supplement 1), S102–S105. Fargione, J.; Hill, J.; Tilman, D.; Polasky, S.; Hawthorne, P. Land clearing and biofuel carbon debt. Science 2008, 319, 1235–1238. Searchinger, T.; Heimlich, R.; Houghton, R. A.; Dong, F.; Elobeid, A.; Fabiosa, J.; Tokgoz, S.; Hayes, D.; Yu, T.-H. Use of U. S. croplands for biofuels increases greenhouse gases through emissions from land-use change. Science 2008, 319, 1238–1240. Smil, V. Enriching the Earth: Fritz Haber, Carl Bosch, and the Transformation of World Food Production; MIT Press: Cambridge, MA, 2001. Galloway, J. N.; Aber, J. D.; Erisman, J. W.; Seitzinger, S. P.; Howarth, R. W.; Cowling, E. B.; Cosby, B. J. The nitrogen cascade. BioScience 2003, 53 (4), 341–356. Fixen, P. E.; West, F. B. Nitrogen fertilizers: Meeting contemporary challenges. AMBIO 2002, 31 (2), 169–176. Gitay, H.; Suarez, A.; Watson, R. T.; Dokken, D. J. Climate Change and Biodiversity; International Panel on Climate Change: 2002; Wageningen University and Research Center Publications: Wageningen, Netherlands; pp 1-86 (ISBN: 92-9169-104-7). Worrell, E.; Phylipsen, D.; Einstein, D.; Martin, N. Energy Use and Energy Intensity of the U.S. Chemical Industry; Lawrence Berkeley National Laboratory: Berkeley, CA, 2000; pp 1-40. Dale, B. E.; Kim, S. In Greenhouse Gas Emission Profile of Ethanol Production under Various Cropping Systems; American Institute of Chemical Engineers: San Francisco, CA, 2003. Bare, J. C.; Gloria, T. P. Critical analysis of the mathematical relationships and comprehensiveness of life cycle impact assessment approaches. Environ. Sci. Technol. 2006, 40 (4), 1104–1113. Belton, V.; Stewart, T. J. Multiple Criteria Decision Analysis: An Integrated Approach, 2nd ed.; Springer: Norwell, MA, 2001; p 372. Boehmel, C.; Lewandowski, I.; Claupein, W. Comparing annual and perennial energy cropping systems with different management intensities. Agric. Syst. 2008, 96 (1-3), 224–236. Urquiaga, S.; Cruz, K. H. S.; Boddey, R. M. Contribution of nitrogen fixation to sugar cane: Nitrogen-15 and nitrogenbalance estimates. Soil Sci. Soc. Am. J. 1992, 56, 105-114. Boddey, R. M.; Urquiaga, S.; Reis, V.; Dobereiner, J. Biological nitrogen fixation associated with sugar cane. Plant Soil 1991, 137 (1), 111–117. Shock, C. C.; Seddigh, M.; Saunders, L. D.; Stieber, T. D.; Miller, J. G. Sugarbeet nitrogen uptake and performance following heavily fertilized onion. Agron. J. 2000, 92, 10-15.

(37) Binkley, D.; Carter, R.; Allen, H. L., Nitrogen fertilization practices in forestry. In Nitrogen Fertilization in the Environment; Bacon, P. E., Ed.; Marcel Dekker, Inc: New York, 1995; pp 421-441. (38) Howarth, R. W.; Boyer, E. W.; Pabich, W. J.; Galloway, J. N. Nitrogen use in the United States from 1961-2000 and potential future trends. AMBIO 2002, 31 (2), 88–96. (39) Sheehan, J.; Camobreco, V.; Duffield, J.; Graboski, M.; Shapouri, H. Life Cycle Inventory of Biodiesel and Petroleum Diesel for Use in an Urban Bus; Final Report NREL/SR-580-24089; National Renewable Energy Laboratory: Golden, CO, 1998. (40) Chisti, Y. Biodiesel from microalgae. Biotechnol. Adv. 2007, 25, 294–306. (41) Wilkie, A. C.; Mulbry, W. W. Recovery of dairy manure nutrients by benthic freshwater algae. Bioresour. Technol. 2002, 84 (1), 81–91.

(42) Tilman, D.; Hill, J.; Lehman, C. Carbon-negative biofuels from low-input high-diversity grassland biomass. Science 2006, 314 (5805), 1598–1600. (43) Tilman, D.; Socolow, R.; Foley, J. A.; Hill, J.; Larson, E.; Lynd, L.; Pacala, S.; Reilly, J.; Searchinger, T.; Somerville, C.; Williams, R. Beneficial biofuelssThe food, energy, and environment trilemma. Science 2009, 325 (5938), 270–271. (44) Lindeijer, E. Biodiversity and life support impacts of land use in LCA. J. Cleaner Prod. 2000, 8 (4), 313–319. (45) Yoder, J.; Young, D.; Painter, K.; Chen, J.; Davenport, J.; Galinato, S. Potential for a Sugar Beet Ethanol Industry in Washington State; Washington State University: Pullman, WA, 2009.

ES902405A

VOL. 44, NO. 10, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3939