Introduction
Fossil fuel energy use and the resulting greenhouse gas (GHG) emissions from food production, transport and consumption contribute significantly to global warmingReference Steinfeld, Gerber, Wassenaar, Castel, Rosales and de Haan1. However, agriculture may contribute to the reduction in GHG emissions by replacing fossil fuel energy used for transport, heating and electricity production with biomass from perennial crops such as grassReference Ney and Schnoor2, willowReference Heller, Keoleian and Volk3, coppiceReference Jørgensen, Dalgaard and Kristensen4 or seasonal crops (e.g. rapeseed, maize)Reference Hanegraf, Biewinga and Van der Bijl5–7. Many of these crops are efficient in terms of reducing GHG emissions, but the studies also show that the total environmental effect of substituting fossil fuel with bio-energy depends on a systems analysis (e.g. whether the indirect energy use of producing the biomass is included). The indirect energy and GHG emissions embodied in inputs like fertilizer are significant in the CO2-emission balances. Therefore, an environmental assessment of bio-energy should include the direct use of energy on farm and the embodied (indirect) energy in inputs used in farm biomass production for electricity and biofuelsReference Ney and Schnoor2, Reference Heller, Keoleian and Volk3, Reference Kim and Dale6.
In response to the European Union's goal of increasing the contribution of biomass to overall energy supply the European Environmental Agency (EEA) analyzed the potential for supplying environmentally compatible bio-energy from agriculture7. It was concluded that agriculture may contribute up to 47 million tons of oil equivalents, which would account for one-third of the target in the EU Commission's Biomass Action Plan (BAP). A significant part of biomass energy should come from the cultivation of set-aside land while EEA assumed that cuttings from grassland would contribute 6–7%. However, the EEA stresses that ‘in order to prevent environmental pressure from the agricultural sector due to more intensive farming [the] study assumed that there will be a high share of environmentally-oriented farming with lower yields’. Organic agriculture is considered as one type of environmentally friendly farming systemReference Dabbert, Häring and Zanoli8, Reference Scialabba and Hattam9, which combines lower emissions of nutrients per hectare with the maintenance of biodiversityReference Stoltze, Priorr, Häring and Dabbert10, Reference Hansen, Kristensen, Grand, Høgh-Jensen, Simmelsgaard and Olesen11. Thus, organic agriculture might play a role in the future bio-energy supply. However, the question of nutrient supply needs to be solved as organic crop rotations with a high percentage of seasonal crops often result in comparably low yields.
The original principles of organic agriculture included goals of reducing dependence on fossil fuel and other limited resourcesReference Woodward and Vogtman12 and this is still an explicit part of the objectives of the Danish organic farming (OF) movement. Due to, among others, the non-use of synthetic fertilizers, energy use is usually lower in OF compared with conventional farming methods, both per hectare and per unit of crop or livestock producedReference Pimentel, Berardi and Fast13–Reference Grönroos, Seppälä, Voutilainen, Seuri and Koikkalainen17. Typically, the fossil energy use per hectare in organic crops is lower compared with conventional crops but yields and stocking density are also lower.
The difference in the use of fossil fuel energy per kg of product between OF and conventional farming systems is often marginal and many OF systems still depend on import of non-renewable energy sources (diesel and electricity). The question of reducing the dependence on fossil fuels in organic agriculture has not been addressed extensively, nor has the potential for organic agriculture to contribute to bio-energy production for society. Jørgensen et al. Reference Jørgensen, Dalgaard and Kristensen4 discuss the potential for using alder (Alnus spp.) as an energy crop in organic farms. Alder may bind atmospheric N2 through symbiotic biological N fixation and can be used for combined heat and electricity production. This perennial crop may also offer other societal benefits such as protection of ground water resources, receiving municipal wastewater and/or be combined with high welfare livestock keeping with intensive livestock production. Fredriksson et al. Reference Fredriksson, Baky, Bernesson, Nordberg, Norén and Hansson18 compared three scenarios for self-sufficiency in on-farm produced bio-based tractor fuels and found that rape methyl ester (RME) had the better energy efficiency compared with ethanol produced from wheat and biogas from leys. Producing and using biogas for traction would require unrealistically large, high-pressure storage facilities on the farm to meet peak season demand. The authors recommended that biogas be used for electricity production instead.
Given this background it is interesting to explore the possibilities for changing organic farms into net-energy producing units by combining food and energy productionReference Jørgensen and Dalgaard19. OF may develop energy-sustainable agricultural systems including bio-energy crops for environmentally sensitive areas and at the same time improve the crop rotations on cash crop farms. This paper will focus on grass–clover leys for biogas-based electricity and rapeseed oil as replacement for tractor diesel.
Methodology
The study consists of three steps: (1) establishing a representative model of Danish organic cash crop farms, (2) modeling the crop yield and energy conversion from introducing bio-energy crops in the rotation and (3) assessing the total energy and nutrient balances at the farm level.
An average organic cash crop farm on sandy soil in Denmark was established as the base farm from a representative farm databaseReference Dalgaard, Halberg, Kristensen and Larsen20. The organic cash crop farm type consisted of the 107 accounts from certified organic farms without dairy production on sandy soil, representing 1084 farms with a total area of 42,287 ha in the base year 1999. The acreage in each crop, the crop yield, livestock production, sales and manure import used in the base scenario were averages from the dataset. Minor crops were not included in the model. Other inputs such as the use of electricity, diesel, manure and feed import were modeled from the crop and livestock production as described in Dalgaard et al. Reference Dalgaard, Halberg and Porter15, Reference Dalgaard, Halberg, Kristensen and Larsen20. The feed import was modeled as the standard feed requirements for organic beef cattle less the amount of home grown feed from leys and cerealsReference Møller, Thøgersen, Kjeldsen, Weisbjerg, Søegaard, Hvelplund and Børsting26. The manure import was calculated according to Danish rules for OF, which limits the total application of manure to 140 kg N ha −1, part of which may be of non-organic origin. The amount of manure imported was checked against national statistical dataReference Dalgaard, Halberg, Kristensen and Larsen20.
The average crop rotation in the base scenario (scenario 1) included only 11% grass–clover and 5% pulses (Table 1). The rest of the 39 ha were seeded to cereals, thus reserving only a small proportion of the land to nitrogen fixing crops. The grass–clover leys and the permanent grassland were used for beef cattle production (12 livestock units). The cattle were supplemented by 7.6 tons of imported rapeseed cake. The 9% set-aside land was not used and had traditionally little content of legumes. The average farm imported 4.4 tons of nitrogen with manure from organic dairy farms and conventional pig farms. Nitrogen and phosphorus balances were calculated according to Kristensen et al. Reference Kristensen, Halberg, Nielsen, Dalgaard, Bos, Pflimlin, Aarts and Vertés21 and Halberg et al. Reference Halberg, Kristensen and Kristensen22 Ammonia losses were estimated according to Andersen et al. Reference Andersen, Sommer, Hutchings, Kristensen and Poulsen23 and denitrification followed Vinther and Hansen.Reference Vinther and Hansen24 The nitrous oxide and methane emissions were calculated according to IPCC25, but using a country-specific accounting method for some of the crop residue N contentReference Møller, Thøgersen, Kjeldsen, Weisbjerg, Søegaard, Hvelplund and Børsting26. The net change in soil N content was estimated using the model C-toolReference Petersen and Berntsen27 and potential nitrate leaching was estimated as the difference in N-surplus, gaseous losses and net soil-N changeReference Knudsen, Kristensen, Berntsen, Petersen and Kristensen28.
Area and yield of crops for three organic farm scenariosFootnote 1 representing an average Danish cash crop rotation and two crop rotations with bio-energy crops.
1 Base, rapeseed and grass–clover denote farm scenarios.
2 Average yield and area of crops estimated from organic cash crop farm database on sandy soil in Denmark.
3 One feed unit (FE) corresponds to the energy value of 1 kg cereal for cattle (~1 kg DM grass–clover and 1.2 kg DM whole crop barley silage).
The crop rotation in scenario 1 is not ideal from an agronomical point of view, because of the relatively weak nutrient re-cycling and the high proportion of cereals, which is also the reason for the dependence on imported conventional manure. Low soil fertility, especially on sandy soils, and limited nutrient supply on organic farms are linked with low crop yields and high weed pressureReference Halberg and Kristensen29–Reference Rasmussen, Askegaard, Olesen and Kristensen31.
Two alternative scenarios were simulated aiming at improved energy self-reliance in the organic cash crop farms. One alternative consisted of producing biofuel for traction from rapeseed on the farm using a low-technology method suitable for individual farms (scenario 2). Rapeseed oil can replace tractor diesel after a slight modification of the engineReference Ansø and Bugge32 by using a farm scale cold press system that extracts 30% of the oil contentReference Ferchau33, Reference Lunderskov34. According to BjergReference Bjerg35 the energy content was estimated to be 35 MJ kg−1 oil corresponding to 32 MJ liter−136. Because rapeseed is a difficult crop to grow in organic rotations in Denmark due to a high nitrogen demand and risk of crop loss due to insects and soil borne diseases, the area with this crop should not exceed 10% of the crop rotation. The yield of the 3.9 ha winter-sown rapeseed was estimated at 1760 kg ha−1 from the data (Table 1). The rapeseed cake (70% of rapeseed weight) after oil pressing was used as cattle feed and thus replaced most of the purchased rapeseed cake (Table 2). The manure import was identical to the baseline scenario.
Inputs used and outputs produced in three scenariosFootnote 1 for Danish organic cash crop farms with bio-energy crops.
1 Base, rapeseed and grass–clover denote farm scenarios.
2 Relatively large amount of oil corresponding to older machines.
3 Returned nutrients from the biogas plant, see text for explanation.
In a third scenario the potential gain from biogas production and recycling of nutrients from an extra 10% grass–clover leys in the crop rotation was estimated. This model included a medium-sized biogas plant within 25 km from the farm, transporting the green mass from the fields in a 28 ton truck and returning (in another truck) the nutrient-rich effluent after digestion. In total 22.4 tons of dry matter (DH) was removed from the grass–clover leys for biogas and 2.3 tons N returned from the biogas plants replacing the imported pig slurry as fertilizer in the cereal crops (Table 2). Thus, the crop rotation in scenario 3 did not depend on non-organic manure because of increased area with biological nitrogen fixation and improved recirculation of nutrients compared with scenario 1. The methane (CH4) biogas yield from grass–clover digestion was estimated at 0.35 kg CH4 kg−1 organic DMReference Sommer, Møller and Petersen37, which is comparable with Fredriksson et al. Reference Fredriksson, Baky, Bernesson, Nordberg, Norén and Hansson18 A net loss of 3% of the methane was subtracted (and included in GHG emissions) with the remainder converted on-site to electricity, which was supplied to the public net. It was assumed that the additional heat generated in the generator was utilized for either local feed processing or heating of houses in a nearby village, which is a typical situation for Danish biogas plants. An estimated 5% of the electricity and 10% of the heat generated was used for pre-heating and handling of grass–clover and was thus omitted from the net-energy calculation.
In order to get conservative estimates of bio-energy production the crop yields in the scenarios 2 and 3 were assumed to be identical to the base model (scenario 1) even though the improved crop rotations would suggest (small) yield increases (see Discussion). The area for bio-energy crops was primarily taken from the set-aside land, subsequently the spring barley area, which was only slightly reduced compared with the base model (Table 1). Therefore, the total production of cash crops and meat were almost identical in the three scenarios (Table 2).
The energy balances and the GHG emissions in the base model and bio-energy scenarios were estimated using the principles of Life Cycle Assessment (LCA) methodology38, Reference Wenzel, Hauschild and Alting39 and coefficients from Nielsen et al. Reference Nielsen, Nielsen, Weidema, Dalgaard and Halberg40 and Dalgaard et al. Reference Dalgaard, Halberg, Kristensen and Larsen20 The inputs of fossil energy in the three models are shown in Table 2. The estimates of direct energy use included all field operations and on-farm transport, transport to and from the biogas plant and transport of imported manure and biogas effluent. Moreover, GHE emissions from the production of farm inputs such as electricity and rapeseed cake were also included in the inventory.
In LCA the resource use and environmental impact aggregated across the product chain are calculated per functional unit, such as kg cereal or beef produced. In order to avoid unnecessary complications the energy use and emissions in the farm models were not allocated among the main products from the farm. However, the farm unit in the three scenarios produced the same total output of 58 tons of cereal, 6 tons of peas and 9 tons of beef (live weight), which facilitates a comparison of energy balances and environmental effects between the models. The small reduction in cereal area in scenarios 2 and 3 was compensated by importing an equivalent amount of organic barley (Table 2). The energy use and environmental impacts from this production outside the cash crop farms were included using data from Nielsen et al. Reference Nielsen, Nielsen, Weidema, Dalgaard and Halberg40 and Dalgaard et al. Reference Dalgaard, Halberg, Kristensen and Larsen20Figure 1 illustrates the systems delimitation for the LCA.
Systems delineation of inputs and outputs in the LCA of three scenarios based on an average Danish organic cash crop farm.
When calculating the net energy balances and the total emission of GHG from the three scenarios it was necessary to account for the saved production of electricity and heat when using the biogas for combined electricity and heat production. This was estimated based on assumptions that a certain part of the Danish electricity production based on fossil fuels was replaced by the bio-energy. A standard value of 9.5 MJ kWh−1, was used, which represents non-optimal fossil electricity production in DenmarkReference Dalgaard, Halberg and Porter15. However, electricity in Denmark is often produced in a combined gas driven power plant, where the heat is used for houses. Therefore, in the alternative calculation, ‘Grass clover II’, a more efficient electricity production system using only 5 MJ kWh−1 was assumed. This reduced both the energy cost in MJ per kWh of imported fossil electricity and the saved energy costs from the electricity exported from the biogas plant. Thus, the fossil energy cost for electricity used on the farm was assumed to be the same as for the replaced electricity production.
The use of crops for bio-energy was considered to be CO2-neutral, which is a standard methodologyReference Ney and Schnoor2, Reference Hanegraf, Biewinga and Van der Bijl5, Reference Fredriksson, Baky, Bernesson, Nordberg, Norén and Hansson18. Thus, the CO2 emission from the combustion of the rape oil or digestion of the grass–clover was not larger than the CO2-sequestration during crop growth. The change in soil C content was not included in the CO2 balances. Methane and nitrous oxide emissions were converted to CO2 equivalents (CO2E) by multiplication by 20 and 320, respectively, in accordance with standard methodology in LCAReference Wenzel, Hauschild and Alting39 and in climate change inventories25.
Results and Discussion
Energy balances
Table 3 shows the energy balance of the average farm before and after the modeled alternative crop rotations. OF uses between 90 and 110 liters of diesel per hectare for field operations, accounting for 59% of the total energy used in the average cash crop farm (base scenario, Table 3). An average rapeseed yield of 1800 kg ha−1 would be enough to replace 50–60% of the diesel used on the farm. The rapeseed cake co-product would replace 64% of the input of concentrate for the farm's livestock production but the indirect energy use saved from this was counterbalanced by the increased electricity needed for the pressing process (Table 3). The fuel/feed use of the rapeseed crop without industrial processing may be the reason for our positive evaluation in contradiction to the assessment of conventional rapeseed for biofuels by Hanegraf et al. Reference Hanegraf, Biewinga and Van der Bijl5 To replace the total diesel use on the average organic cash crop farm would require either a yield of 2800 kg rapeseed per hectare on the 10% of the land or greater oil extraction rates, which may in fact not be unrealistic in winter-sown rapeseed in the future. Another option for utilizing the energy yield from rapeseed exists, namely the conversion of oil to RME. Fredriksson et al. Reference Fredriksson, Baky, Bernesson, Nordberg, Norén and Hansson18 found that an organic cash crop farm would be self-sufficient in bio-fuel production if 9.3% of the land was grown with rapeseed for RME. They assumed a yield of 2000 kg rapeseed per ha and 68% oil extraction efficiency which resulted in a net energy production in fuel of 18.5 GJ ha−1 compared with 14.6 GJ ha−1 in our study. However, the transesterification in itself does not yield more energy from rapeseed oil and the process depends on electricity and external chemical inputs including a supplement of methanol. Therefore, we consider the use of cold pressed oil directly in the tractor as more in line with organic philosophy. Fredriksson et al. Reference Fredriksson, Baky, Bernesson, Nordberg, Norén and Hansson18 also found that RME would be more efficient compared with wheat-based bioethanol and biogas as home produced biofuel in OF and conclude that biogas is more suitable for electricity production, which agrees with our study.
Energy use and productionFootnote 1 (GJ yr−1) for three scenarios for Danish organic cash crop farms with bio-energy crops.
1 All farms export 58 tons of cereals, 6 tons of pulses and 9 tons of beef (live weight).
2 The fossil energy cost for electricity used on the farms is assumed to be the same as for the replaced electricity production, see Methods section for details.
3 A small import of cereals was included in the models to compensate for the reduced barley area, so that all farms deliver the same amount of products.
4 Assuming that the heat generated from the electricity production is used efficiently, which is questionable—see text for discussion.
5 Base, rapeseed and grass–clover denote farm scenarios.
Increasing the area with grass–clover from 10 to 20% of the crop rotation (in the scenario 3) gives a surplus of DM, which could produce 7840 m3 of biogas. When using the biogas for a combined generation of electricity and heat, the grass–clover yielded more than 2.5 times the electricity used on the cash crop farm (247 GJ versus 90 GJ, Table 3). This corresponds to approximately 2600 kWh from 3.9 ha grass or enough to cover the electricity in 5–6 Danish households. The generated 111 GJ of heat (as a by-product of electricity production from the biogas) may be used for producing feed-mixtures, evaporating water from the degassed grass effluent or for heating houses in a village through pipelines (a type of centralized heating system very often used in Denmark already). Assuming that all electricity and heat is utilized for purposes where it saves fossil fuel energy, the farm will be a net energy producer (also after deduction of the 15 GJ energy used for transport of grass and de-gassed grass-effluent to and from the central biogas plant, Table 3). Contrary to the rapeseed oil this energy was not produced nor used on the farm, but farm scale biogas production is a possibilityReference Fredriksson, Baky, Bernesson, Nordberg, Norén and Hansson18.
The conversion between kWh and GJ depends on how efficient the marginal electricity production is (i.e. the type of fossil energy saved when bioelectricity is generated). This assumption was tested with different values as shown in the scenarios 3a and 3b in Table 3. If the electricity replaced was of a more efficient type, then the net energy gain from the farm would be 53 GJ or 28% of the energy used on the farm. From Table 3 it also appears that if only 50% of the generated heat was utilized then the net energy gain in the grass–clover scenarios would be reduced to 72 and 0 GJ, respectively. However, it would be agronomically possible to combine the rapeseed and grass–clover energy crops into one scenario, in which case the farm would be self-sufficient with energy. But such a scenario would then use 19% of the farm's land for energy production, which would reduce total yield of food crops. The economical and political implications of such a change in agricultural land use is beyond the scope of this study, but does seem to be within scenarios considered politically acceptable in the EU7.
Emission of GHG
Table 4 shows the estimated GHG emissions in CO2E. Nitrous oxide emitted from the farms contributed the larger part (69–77%) of total on-farm GHG emissions and equals 82 tons CO2E for all three scenarios. Denitrification of leached nitrate accounted for 26–35% of nitrous oxide emissions. Less nitrate (46 kg N ha−1) was leached from the ‘grass–clover scenario’ compared to the other two scenarios (both 61 kg N ha−1) (Table 5). Also less N2O (22 kg CO2E) from denitrification of nitrate was emitted from the ‘grass–clover scenario’ compared to the other scenarios (both 29 kg CO2E). However, the N2O emissions from biological nitrogen fixation and crop residues were higher on the grass–clover scenario (29 kg CO2E) compared to the other scenario (both 22 kg CO2E), thus the total nitrous oxide emissions and nitrate leaching in the ‘grass–clover scenario’ was counterbalanced by higher biological nitrogen fixation and crop residues (Table 4).
Emissions of GHG for the three scenariosFootnote 1 for Danish organic cash crop farms with bio-energy crops.
1 Base, rapeseed and grass–clover denote farm scenarios.
2 See Footnote note 2 in Table 3.
3 See Footnote note 4 in Table 3.
Field-level nitrogen balances (kg N ha−1) of three scenariosFootnote 1 for Danish organic cash crop farms with bio-energy production.
1 Base, rapeseed and grass–clover denote farm scenarios 1, 2 and 3.
2 N imported in slurry from pig and dairy farms. In the ‘grass–clover’ scenario the N in residues from biogas plant replaced an equivalent amount of N in imported manure (mostly conventional pig slurry).
The effect of biogas production was a reduction of total GHG emissions from 115 to 92 tons CO2E (20%) if assuming unchanged nitrous oxide emissions and including the effect of replaced fossil energy use for electricity. Further studies are needed to determine the degree of uncertainty in these estimates.
The CO2 balances from crop production and bio-energy use in the different scenarios were not included in the GHG emissions and were treated as neutral. This probably underestimated the net improvements in CO2 balances of the bio-energy systems compared with the fossil fuel system because the CO2-sequestration in the roots and stubble of the grass–clover crop will contribute to a (higher) net build up of soil-carbon, which is not emitted in the bio-energy production. The net soil organic matter gain in the different rotations was not included directly in the calculations of GHG emissions. However, using the calculated average net soil-N gain and a C/N ratio of 10, the extra C sequestration due to increased grass–clover area in scenario 3 amounts to approximately 100 kg C ha−1 compared with scenarios 1 and 2 (Table 5). This is equivalent to 367 kg CO2 sequestered ha−1 y−1 or approximately 14.3 tons CO2 on the 39 ha, which corresponds to a further 12% reduction in GHG emissions compared with scenario 1 (Table 4).
Nutrient balances and losses
Table 5 shows the nutrient balances and losses in the different crop rotations. There was no detectable difference between the base and the rapeseed scenario. Harvesting grass–clover for biogas and returning nutrients in the ‘grass-effluent’ to the cash crops (scenario 3) improved nitrogen cycling on the farm and reduced the need for imported conventional manure. The returned residues from degassed grass–clover thus replaced the imported conventional pig slurry in the grass–clover scenario. Because the biological nitrogen fixation was higher, the N input in scenario 3 was 9% higher compared to the base crop rotation (scenario 1). However, the export of N with grass–clover resulted in a 34% higher output and a slightly lower farm gate N-surplus (96 versus 101 kg N ha−1). The potential leaching was reduced from 61 to 46 kg N ha−1 because the increased grass–clover area led to an increase in the average net soil-N gain through increased soil organic matter build-up.
As mentioned, the crop yields per ha were kept constant across the scenarios. However, in reality the improved crop rotation, nutrient recycling and the soil fertility improvement from grass–clover would probably increase the cereal yields and decrease weed pressure and diseases. In a parallel studyReference Dalgaard, Olesen, Halberg and Berntsen42 with the model FASSETReference Berntsen, Petersen, Jacobsen, Olesen and Hutchings41 an increased proportion of grass–clover in an organic cash crop rotation improved soil fertility and nutrient cycling and resulted in approximately 10% higher cereal yields. The EEA states that the use of grassland cuttings for energy purposes may be a good opportunity to maintain the management of extensive farmland in order to preserve biodiversity (as an alternative to abandonment)7. In our study we assumed that the 3.3 ha permanent grassland would be used for grazing beef cattle. But in case of reduced grass needs on the farm, the permanent grassland could yield an additional 5–6 tons of DM for biogas production or a 25% increase compared to the grass–clover scenario.
The rapeseed and grass–clover crops were used as examples of the potential for energy self-reliance in OF. Even though they were modeled separately for simplicity, both could be implemented at the same time and complement each other in order to maximize energy self-reliance. Other crops such as coppice may prove to be more profitable and require less land in a long-term perspectiveReference Jørgensen, Dalgaard and Kristensen4, and it is eventually a normative question for the organic movement whether bio-energy should be produced and used on the farm.
Conclusion
Changing average organic cash crop rotations in Denmark to include 20% grass–clover and 10% rapeseed could serve the dual purpose of building soil fertility and delivering biomass for energy production for society or for energy self-sufficiency at the farm level. While this possibility has become more attractive at the farm level after the reform of the EU Common Agricultural Policy, there are a number of structural and organizational challenges to overcome, such as locating biogas plants at economic hauling distances from a concentration of organic farms and near a facility that may use the surplus heat produced.
