Introduction
Increased public awareness of the need for actions against global warming has brought greenhouse gas (GHG) emissions and the use of fossil energy into focus in agricultural organizations in general and organic agriculture in particularReference Scialabba and Müller-Lindenhauf 1 . The reduction of external inputs to the farm through efficient management of materials and energy is part of the organic principles established by the International Federation of Organic Agriculture Movements (IFOAM) 2 . Also in the regulation of organic agriculture in Europe, it is stated that organic farming should be based on ‘a minimisation of the use of non-renewable resources’ [Council Regulation (EC) No. 834/2007]. Reduction in the dependency on fossil fuels has been identified as one of the important areas for development of organic farming in Denmark 3 . Another important issue is the nitrogen (N) supply that is currently provided in part by the import of animal manure from conventional farms and partly by the use of N-fixing crops as green manure. The Danish organic farmers association has planned to phase out use of animal manure from conventional farms by 2021, in order to make organic agriculture independent of conventional farming systems 4 .
In 2010, the total area grown organically in Denmark was 173,513 ha, corresponding to 6.6% of the agricultural land 5 . In Denmark, the trade of organic products was doubled between 2005 and 2008, and was until 2011 still increasing by approximately 5% (by value) per year 6 . A further increased conversion of land is necessary if Danish products are to contribute to a continued growth in the organic sector 3 , and a doubling of the current organic farm area is planned, together with a significant increase in the production of biogasReference Dalgaard, Kjeldsen, Kristensen, Kristensen, Alrøe and Halberg 7 . At the same time there is a need for improvement of the production system to meet the energy and climate challenge described above, without compromising other organic principles.
Fossil energy use is generally lower in organic than in conventional agriculture. A 41–65% reduction in energy use per hectare for production of winter wheatReference Gomiero, Paoletti and Pimentel 8 and a 30–35% reduction for production of spring-sown cereals have been reportedReference Refsgaard, Halberg and Kristensen 9 . When energy use efficiencies (energy use per unit produce) are compared, the difference is smaller and numbers from 33% less to 21% higher energy use for grain production have been reportedReference Gomiero, Paoletti and Pimentel 8 – Reference Dalgaard, Halberg and Porter 10 . When yields in organic farming are substantially lower than in conventional farming, energy efficiencies drop and approach, or even surpass, those of conventional farming. For fodder crops such as grass–clover, with high mineral fertilizer application rates in the conventional system and relatively high yields in both systems, the energy efficiency is, however, up to 90% higher in the organic than in the conventional systemReference Gomiero, Paoletti and Pimentel 8 , Reference Dalgaard, Halberg and Porter 10 .
Halberg et al.Reference Halberg, Dalgaard, Olesen and Dalgaard 11 explored options for energy self-sufficiency on organic farms and compared energy production from grass–clover-based biogas to energy production from rapeseed. The biogas scenario was superior to rapeseed on energy balance and furthermore reduced both GHG emission and N leaching. Anaerobic digestion of grass–clover cuttings also provides the possibility for production of organic fertilizer on stockless organic farms, when the digestate (residual product from anaerobic digestion) is recycled to the fieldReference Stinner, Möller and Leithold 12 . This decreased dependency on animal manure as a fertilizer might provide the possibility of expanding the area of organic cash crop productionReference Tersbøl, Alrøe and Halberg 13 . When compared with a traditional system, where grass–clover is used as green manure, use of digestate as a fertilizer provides a more even allocation of nutrients in the crop rotation, a better matching of N demand and supply, and thereby higher yieldsReference Stinner, Möller and Leithold 12 . Furthermore anaerobic digestion of slurry on livestock farms has the potential of reducing the emissions of methane (CH4).Reference Sommer, Petersen and Møller 14
It has been suggested that a supplement of nutrients in organic farming could be achieved by use of biomass from extensively managed or non-used grasslands. This would in addition contribute to the maintenance of meadows and permanent grasslands, where cutting is essential in order to remove nutrients and maintain biodiversityReference Nielsen, Hald, Hopkis, Gustafsson and Bertilsson 15 .
The implementation of biogas production on organic farms and its effects depend on the existing farming system and how the implementation changes the farming system. Therefore the aim of this study was to analyze (1) how biogas production in organic farming may be achieved through changes in the crop rotation, and (2) how different scenarios for biogas production affect productivity, N flows, energy balance and GHG emissions of the farm.
Methods
The study applied four steps: (1) establishment of a 1000 ha organic farm model representing organic agriculture in Denmark (reference scenario), (2) creation of scenarios for biogas production on the farm, and (3) analysis of environmental impacts focusing on N flows, crop yields, net energy production and GHG emissions. An updated version of the whole-farm model FarmGHGReference Olesen, Schelde, Weiske, Weisbjerg, Asman and Djurhuus 16 , Reference Olesen, Weiske, Asman, Weisbjerg, Djurhuus and Schelde 17 was used for the analysis. FarmGHG is a tool designed for quantification of carbon (C) and N budgets in dairy farm systems and for quantification of direct and indirect gaseous emissions from dairy farms (see section ‘Environmental impacts’).
Establishment of a 1000 ha organic farm model
Land use
The establishment of a representative 1000 ha organic farm was based on analysis of datasets from the National Danish Agricultural Register. This register contains information on land use in Danish agriculture and, for organic farmers, information on the number of livestock on the farm. The information is derived from EU Single Payment applications of Danish farmers, which are collected in a digital database where all farmers are obliged to declare crop type areas and livestock numbers in order to receive subsidies. These data were combined with data on fertilizer accounts from the Danish Plant Directorate and with soil types as described by Dalgaard et al.Reference Dalgaard, Rygnestad, Jensen and Larsen 18 . From both registers data from 2006 were used. The combined datasets of 2622 farmers having solely organically farmed areas in their application were divided into groups depending on working hours, livestock and soil type, following Kristensen and KristensenReference Kristensen, Kristensen and Brebbia 19 . Full-time dairy farms and full-time cash crop farms on sandy soils (soils containing less than 15% clay and less than 30% silt) were chosen for further analysis. These two farm types represent 30% of organic farms in Denmark, but manage 61% of the organically certified land. Data on land use, number of livestock and use of manure were averaged for comparative analysis (Figs 1a and 1b, Table 1). Woodland and minor crops representing less than 1% of the area on both farm types, and livestock species other than dairy cows and heifers were not included. The two farm types were combined according to the actual distribution of organic dairy farms and cash crop farms on sandy soil; 32% of the area was managed as cash crop farms and 68% of the area was managed as dairy farms (Fig. 1c, Table 1). The average import of slurry from conventional farmers corresponded to 26 kg N ha−1 yr−1. The combined farm formed two reference scenarios used for the further analysis: Reference A represented the current situation in Danish organic agriculture with import of manure in slurry from conventional farms, and Reference B represented a situation without import of conventional slurry.
Figure 1. Land use on (a) Danish cash crop farms on sandy soil and (b) Danish dairy farms on sandy soil. (c) The area use of dairy farms and cash crop farms is combined to establish a 1000 ha farm model used as reference scenarios (References A and B).
Table 1. Average area and number of livestock on organic cash crop farms and dairy farms on sandy soil in Denmark.
1 Heifers on cash crop farms are not scaled up to the combined farm.
Livestock
Milk yield followed the Danish annual farm account statistics for 2006 20 . The feed requirement was calculated for cows and heifers according to Olesen et al.Reference Olesen, Weiske, Asman, Weisbjerg, Djurhuus and Schelde 17 . A feed plan was developed (Table 2) based on the crop distribution. It was assumed that the 1000 ha combined farm was self-sufficient with grass–clover and roughage. Feed import was reduced to a minimum and consisted exclusively of barley grain. Usually organic dairy farms import concentrate with a higher protein content than cereal grain 21 , but high amounts of protein in the grass–clover compensated for this and the protein content in the diet for the dairy cows followed Danish standardsReference Poulsen, Børsting, Rom and Sommer 22 .
Table 2. Livestock numbers, milk yield and feed plans for cows and heifers.
* 1 SFU corresponds to 12 MJ metabolizable energy (equivalent to 1 kg of barley).
The National Danish Agricultural Register does not contain information on whether the grass–clover area is used for grazing, silage or other purposes, e.g., green manure. Therefore, the area used for grazing and silage was estimated from the need for roughage, and the remaining area was used for green manure. Organic cattle must have access to grazing from April 15 to November 1, whenever weather and the physical condition of the cattle allow 23 . Based on data from the Danish Agricultural Advisory Service (DAAS) 21 it was assumed that the dairy cows were grazing 8 h per day and heifers were grazing 24 h per day in 184 days out of the 199-day potential grazing season. In the grazing season, it was assumed that the cows consumed 7 Scandinavian feed unit (SFU) per day from grazing. 1 SFU corresponds to 12 MJ metabolizable energy (equivalent to 1 kg of barley). The heifers covered 100% of their feed requirement from grazing in the grazing season.
Manure
The N demand of the crops was determined using the Danish fertilizer norms for each crop 24 (Table 3). The fertilizer norm is currently set to 85% of the gross margin based, farm economical optimum of plant available N for each crop. If the previous crop is an N-fixing crop, it contributes to plant-available N and reduces the fertilizer norm. Therefore, it was necessary to define a simplified crop rotation. It was assumed that grass–clover fields were rotated every second year. Grass–clover was followed by spring wheat, oats, maize and potatoes. The manure available was insufficient to fulfill the N norm. Therefore, the N application was reduced equally for all crops. NH4-N applied corresponded to 46% of the N norm (Table 3). Crop yields followed the DAAS estimates for organic crops on sandy soil 25 (Table 4). Maize yield and N application were unchanged when maize was grown for bioenergy. However, N application for grass–clover as an energy crop was reduced to the same level as for grass–clover used for grazing. This was based on the assumption that grass–clover as an energy crop has a lower priority than grass–clover for roughage. The grass–clover yield was adjusted to the lower N application assuming a N response on 13 kg dm kg N−1 as reported by SøegaardReference Søgaard 26 .
Table 3. Fertilizer norms 24 and distribution of manure in Reference scenario A.
Table 4. Organic crop yields on sandy soil 25 .
Scenarios
The overall aim for implementation of biogas production on organic farms in our study was to achieve self-sufficiency with organic fertilizer and energy. Four scenarios for implementation of biogas production were created in order to cover possible implications of biogas production on organic farms: (S1) use of grass–clover or (S2) maize silage as an energy crop, (S3) extension of the cash crop area, and (S4) withdrawal of biomass from unmanaged meadows through cutting and baling.
In S1 and S2, 100 ha were used for production of energy crops (grass–clover and maize silage) for anaerobic digestion (Table 5). In both scenarios, the area for feed production and the number of livestock were sustained. The 100 ha energy crop was obtained by omitting the grass–clover for green manure and reducing the cash crop area. In S3 and S4, the cash crop area was doubled and 200 ha were used for production of grass–clover for biogas. As a consequence, the area with grass–clover for green manure was omitted and the area for roughage was reduced. The number of livestock on the 1000 ha was reduced proportionally to the decline in roughage area. In S4, an additional 200 ha of extensive meadow (outside of the 1000 ha farm) was harvested and used for biogas production. In all scenarios, the slurry from the livestock excreted in-house was used for biogas production together with the energy crop. The digestate was used as a fertilizer, and the import of manure from conventional farms was omitted. In order to sustain roughage yields and avoid changes in the feed plan, the feed crops received the same amount of plant-available N in the references and in all scenarios. The remaining N was applied to the cash crops. The grass–clover area, and thereby the area of crops having grass–clover as a pre-crop, differed between scenarios. In S1, grass–clover was followed by spring wheat, oats, maize, potatoes and winter wheat, in S2 the crop rotation used in the reference scenarios was used, and in S3 and S4 grass–clover was followed by spring wheat oats and maize.
Table 5. Land use (ha) in References (A + B) and the scenarios for biogas production (S1–S4).
The changed crop rotation and the use of biogas digestate as fertilizer affected the amount and quality of the fertilizer in the scenarios. Therefore, the yield of the cash crops and energy crops was adjusted depending on the amount of ammonia N applied to the individual crops. The yield response used was based on recommendations by DAAS 27 (Table 6). Only yields of cereals were adjusted, since manure application to feed crops was kept constant. Yields of potatoes and other cash crops, which comprise less than 5% of the area, were not adjusted due to lack of data on the N response of these crops in organic farming.
Table 6. Yield adjustments with varying rates of ammonia N applied as slurry 27 .
Anaerobic digester
CH4 production in the FarmGHG model is determined by the equation: PCH4 = ε VS Bo 0.67, where PCH4 is the CH4 production (kg), VS is the input of volatile solids to the digester (kg), Bo is the maximum methane-producing capacity of the added material (m3 kg VS−1), ε is the efficiency of the process (0–1) and 0.67 converts from volume to kg of CH4.Reference Olesen, Weiske, Asman, Weisbjerg, Djurhuus and Schelde 17 Methane-producing capacities of the materials used are showed in Table 7. Reliable values for ε were not found in the literature, but since practical yields generally are lower than Bo,Reference Møller, Nielsen, Murto, Christensson, Rintala, Svensson, Seppälä, Paavola, Angelidaki and Kaparaju 28 an efficiency of 0.9 was assumed.
Table 7. Parameters for anaerobic digestion used in FarmGHG.
The anaerobic digester specified in the FarmGHG model consists of three components: a pre-store tank, a digestion tank and a post-store tank (Table 7). In order to minimize GHG emissions from the tanks, it was assumed that both the pre-store and the post-store had a solid cover. This also reflects the general standard in DenmarkReference Jørgensen 29 . The surface areas of the tanks were calculated from the volume of material added to the respective tanks in S1, assuming the height of the tanks was 6 m. The actual amounts of organic matter in the stores varied between scenarios, but the surface areas were kept constant. Water was added to the digester to reach a dry matter level of 10% to make the material liquid enough for pumpingReference Jørgensen 29 .
The energy requirement of the anaerobic digester was defined based on Berglund and BörjessonReference Berglund and Börjesson 30 . They reported best estimates for continuous, single-stage, mesophilic tank reactors; a common reactor type in DenmarkReference Jørgensen 29 . Heat and electricity inputs of 1100 and 660 MJ per ton dry matter were used.
The lower heating value of CH4 (35.8 MJ m−3) was used to decide the energy production from the CH4 produced in the digester. A total of 35% of the energy produced was assumed to be converted into electricity, 50% was converted to heat, whereas 15% was lostReference Jørgensen 29 . CH4 loss from the anaerobic digester was set to 1.8%. 31
Environmental impacts
The environmental impacts analysis focused on the direct energy balance, the N balance and the GHG emissions of the farm.
Energy use and production
Energy use on the farm included electricity for animal production (400 kWh cow−1 yr−1), for cleaning the floor (40 kWh cow−1 yr−1), and diesel for management of the fields as specified in the FarmGHG modelReference Olesen, Weiske, Asman, Weisbjerg, Djurhuus and Schelde 17 . Indirect energy use for production and provision of goods and services to the farm is not included in the analysis. Energy production on the farm was based on CH4 production in the anaerobic digester. Since the biogas plant was on the farm, no transport to and from the biogas plant was included. However, energy for transport of imported slurry and biomass from meadows was calculated according to Dalgaard et al.Reference Dalgaard, Halberg and Porter 10 and added to the energy use. A transport distance of 6 km in total was assumed.
N balance and N loss
A farm N balance was calculated according to Watson et al.Reference Watson, Bengtsson, Ebbesvik, Loes, Myrbeck, Salomon, Schroder and Stockdale 32 using numbers extracted from the FarmGHG output files. N deposition was not included in the FarmGHG model, but an input of 15 kg N ha−1 (equivalent to the average N depositions in Denmark in 2005Reference Ellermann, Andersen, Bossi, Brandt, Christensen, Frohn, Geels, Kemp, Løfstrøm, Mogensen and Monies 33 ) was added to the N balance. The N input includes: slurry import, import of grass cuttings from meadows, feed, seed, N fixation and deposition. The N output includes: milk, meat and cash crops sold from the farm. The FarmGHG model estimates N losses in the form of nitrate leaching, ammonia volatilization and nitrous oxide emissions. The nitrate leaching is calculated as 30% of all N input to the field. The model SimDenReference Vinter and Hansen 34 was used for quantification of denitrification in the soil. The model calculations are based on the amount of N applied and N source; in this case legumes, slurry and grazing livestock. The denitrification is quantified for various soil types and for three levels of organic matter in the soil. We assumed a medium level of organic matter in the soil, and used the mean of the denitrification rate from the model for the four classes of sandy soils in Denmark. Changes in the soil N pool were estimated by subtracting all N losses from the N surplus.
GHG emissions
The FarmGHG model uses Intergovernmental Panel on Climate Change (IPCC) standards for calculations of GHG emissions. In this case, the default methodology of the model was updated with the IPCC 2006 emission factors 35 .
The GHG emissions calculated included emissions of N2O, CH4 and CO2, both direct emissions from the farm and pre-chain emissions from import of energy, seed and feed. The import of slurry and organic matter from meadows is not associated with any pre-chain emissions, because these products are regarded as waste. CO2 emissions for transportation of imported slurry and biomass from meadows were calculated from the diesel use according to Olesen et al.Reference Olesen, Weiske, Asman, Weisbjerg, Djurhuus and Schelde 17 , and added to the CO2 emissions calculated by FarmGHG. N2O emissions included pre-chain emissions, emissions from slurry in the house and manure store, and emissions from the field. Also indirect N2O emissions from N lost by nitrate leaching and ammonia volatilization were accounted for. The CH4 emissions included pre-chain emissions, emissions from enteric fermentation in the livestock, emissions from slurry in the house, emissions from the manure store and anaerobic digester, and emissions from feces excreted in the field. CO2 emission or sequestration from changes in soil organic matter content is not included in the model estimation, since the farm system is assumed to be in steady stateReference Olesen, Weiske, Asman, Weisbjerg, Djurhuus and Schelde 17 .
Results
N fluxes and yields
The amount of N available in the farm systems depended on the number of livestock, the grass–clover area, the energy crop used, and the import of slurry or organic matter (Table 8). The total amount of N available for application was higher in Reference A, where slurry and feed were imported, than in the four scenarios with use of digestate from the biogas production. In Reference B, it was not possible to sustain the fertilization of the roughage and no N was left for the cash crops. In S3 and S4, the increased amount of digestate from energy crops did not compensate for the lower livestock density and lack of slurry import when the total amount of N for application was considered. Although the total amount of N available for application was lower in the biogas scenarios, it was possible to apply NH4-N to the cash crops in amounts equivalent to (S3) or higher (S1 and S4) than the amounts used in Reference A. Thereby increased cash crop yields were obtained in S1 and S4. Omission of import of manure in Reference B reduced total farm cash crop yield by 19% compared with Reference A. In S1 and S2, where the cash crop area was reduced due to the production of energy crops, the total cash crop yield was 11 and 28% lower than in Reference A.
Table 8. Manure use and crop yields for 1000 ha farms.
Environmental impacts
Energy balance
Energy use, in the form of electricity and diesel, was higher in the biogas scenarios than in References A and B due to the energy requirement of the anaerobic digester (Fig. 2). This was, however, compensated for by the energy production of the anaerobic digester, and all biogas scenarios had a positive energy balance. The net energy production in the four scenarios was between 6274 GJ (S1) and 11,130 GJ (S4). This net energy production assumed that the heat was utilized and not wasted. In S3, where the area used for bioenergy production was doubled and the number of livestock was decreased, the net energy production was increased by 44% compared with S1.
Figure 2. Use of electricity and diesel and production of electricity and heat. Diesel use for transport includes solely the consumption in relation to the external import of slurry and meadow grass.
N balance and N loss
The N surplus of the farm was between 79 and 107 kg N ha−1 (Table 9). N fixation was the largest single contributor to the N surplus. Other important factors affecting the N surplus were the amount of imported slurry or organic matter and the export of milk and cash crops. The export of N from the farm was at the same level in all scenarios; the decrease in export of milk and meat in S3 and S4 was compensated by an increase in export of cash crops. Another result of the larger cash crop area in S3 and S4 was that no import of feed was needed.
Table 9. Farm N balance (kg N ha−1 yr−1).
S2, S3 and S4 reduced the N-leaching loss by 7.7, 16.1 and 11.9%, respectively, compared with Reference A, and S1 increased the loss by 4.8%. N lost by denitrification followed a similar pattern. The NH3 emissions were reduced by 18.9 and 22.5% in S1 and S2, when compared with Reference A. In S3 and S4, where the number of livestock was reduced, the reduction was 45.5 and 40.5%. The calculations showed a risk of decreasing amounts of N in the soil in Reference B and S2.
GHG emissions
The net GHG emission was reduced by the implementation of anaerobic digestion (Fig. 3). The net GHG emission of S1 and S2 was reduced to 64 and 62% of Reference A. In S3 and S4, a large reduction in net GHG emission (to 17 and 12% of Reference A) was mainly caused by reduced CH4 emission from the reduced number of livestock (Fig. 4a). CH4 emissions from the manure store were reduced by the anaerobic digestion, but the loss of CH4 from the biogas plant resolved part of this reduction. Seventy percent (S2) to 96% (S4) of the N2O was emitted from the field (Fig. 4b). The N2O emission from the manure store was slightly higher in the biogas scenarios than in the reference scenarios, partly due to indirect emission caused by larger ammonia emissions. The pre-chain emission of N2O was reduced considerably in S3 and S4, where no feed was imported. Biogas production reduced the indirect emissions of CH4 and N2O connected to transport and production of diesel and electricity, illustrated by a small negative emission. CO2 emission was in all scenarios smaller than the reduction in CO2 emissions from fossil fuels caused by the biogas production (Fig. 4c).
Figure 3. Net emission of N2O, CH4 and CO2 from the 1000 ha farm. The numbers above the columns show the total net GHG emission.
Figure 4. Emissions of (a) CH4, (b) N2O and (c) CO2 for the 1000 ha farm divided into origin and expressed in CO2 equivalents.
Discussion
Manure availability and yields
The current import of slurry from conventional farms to organic farms in Denmark is represented by 25.8 kg total N ha−1 imported in Reference A. The use of 10% of the land for bioenergy production in S1 and S2 does not produce enough digestate to compensate for this import: grass–clover contributes 17.7 kg N ha−1, whereas the contribution of maize is just 11.0 kg N ha−1 (Table 8). However, the annual N fixation of 100 ha grass–clover energy crop in S1 corresponds to 24.9 Mg N and the total N input to the farm is therefore at the same level as in Reference A. Therefore, when the lower cash crop area and the higher fertilizer replacement value of the digestate are taken into consideration, use of grass–clover for biogas increases the amount of N available to the cash crops (Table 8). On the other hand, maize does not provide extra N to the farm, but requires N fertilization to obtain high yields. This causes a lower total input of N to the farm in S2, and the yields of the cash crops are reduced. Based on this, grass–clover seems to be the most realistic option for biogas production in organic agriculture, assuming that the energy crop should contribute both to energy supply and improve N availability on the farm.
In the present study, grass–clover for biogas production was fertilized despite the high amount of N fixed by the clover. Grass–clover yields respond positively to N application, but the N response of grass–clover is variable and dependent on a number of factors, including the share of clover in the field and the age of the grass–clover cropReference Søgaard 26 . Relatively high yields (7.1–7.8 Mg dm yr−1 on average) have been reported without any N applicationReference Søgaard 26 . Therefore, a better strategy for production of grass–clover as an energy crop might be to omit fertilization and spare the N for the cash crops.
Analysis of the geographical distribution of organic dairy and cash crop farms in Denmark shows that the cash crop farms are only to a limited extent placed in the same regions as dairy farmsReference Dalgaard, Kjeldsen, Kristensen, Kristensen, Alrøe and Halberg 7 . Therefore most cash crop farms depend on either import of slurry from conventional farms or on N-fixing crops as a source of N to the crop rotation. In S3, 100 ha grass–clover provides enough N to sustain a doubling of the cash crop area (from 243 to 486 ha). In Reference A, the cash crops receive on average 63 kg total N ha−1, equivalent to 30 kg plant-available N (Table 3). The 100 ha grass–clover field in S1 yielded 17.7 Mg N yr−1. This corresponds to a harvest of 106–123 kg plant-available N ha−1 yr−1 (assuming a content of 60–70% plant-available N in the digestate). Usage of the digestate from 1 ha grass–clover provides enough N for 3.5–4.1 ha cash crops, maintaining the amount of N currently applied in organic agriculture in Denmark. This corresponds to approximately 20% of the farm area. Based on this, biogas production has the potential to make organic cash crop producers independent of conventional manure and thereby contribute to an increase in the organic cash crop area, as suggested by TersbølReference Tersbøl, Alrøe and Halberg 13 . In another study of anaerobic digestion of grass–clover in organic cash crop farming systems by Stinner et al.Reference Stinner, Möller and Leithold 12 , 1/6 of the area was used for grass–clover. Here, the grass–clover was digested in combination with cover crops and crop residues. Thereby more N was recycled in the digestate and on average 125 kg total N ha−1 was supplied to the cash crops, showing that the N potential of the current study is perhaps conservative.
The use of organic matter from meadows in S4 provides an additional option for provision of N, as well as other nutrients, to the system. In a system with export of cash crops, but without external imports (e.g., manure and feed), essential nutrients are removed from the system. In the long term, the risk is that soil nutrient reserves [e.g., phosphorus (P) and potassium (K)] are depleted, as demonstrated by Gosling and ShepherdReference Gosling and Shepard 36 . Another study by Møller and NielsenReference Møller and Nielsen 37 suggested that 100,000 ha meadows are suitable for harvest of biomass for bioenergy production in Denmark; and therefore an area of 200 ha of meadow for a 1000 ha organic farmland is not unrealistic. However, accessibility, yield and interest of the plot owners are important factors that must be investigated to clarify the actual potential.
Environmental impacts
Energy balance
The 1000 ha farm model has an energy surplus, but import of diesel is still necessary to fuel the farm machinery. In that sense energy self-sufficiency is not obtained. In Denmark, biogas is generally used for combined heat and power production, and the heat used for district heatingReference Møller, Nielsen, Murto, Christensson, Rintala, Svensson, Seppälä, Paavola, Angelidaki and Kaparaju 28 . As the heat requirement on farm is limited, it is important to be aware that the placement of the biogas plant is restricted to locations with possible applications of the heat, e.g., urban areas with district heating. This conflicts with the assumption that the biogas plant is placed on the farm. Alternatively the biogas could be upgraded for distribution in the natural gas network in Denmark or used as a gaseous transport fuel, as suggested by Smyth et al.Reference Smyth, Murphy and O'Brian 38 .
The electricity requirement of the farm is very low compared with other investigations, where electricity use on dairy farms 4–5 times higher than the value used in FarmGHG has been reportedReference Gomiero, Paoletti and Pimentel 8 , Reference Refsgaard, Halberg and Kristensen 9 . However, energy companies advise farmers to reduce their electricity use if it exceeds 800 kWh cow−1 yr−1 (Kurt Mortensen, personal comment). The default value used by FarmGHG (440 kWh cow−1 yr−1) is perhaps optimistic, but was used due to a lack of more reliable data.
Energy costs of transport of biomass to and from the digester, apart from the imported biomass from meadows in S4, were neglected as a single farm structure with short distances was assumed. In a study by Gerin et al.Reference Gerin, Vliegen and Jossart 39 , 10 km transport of maize and grass–clover silage to a centralized biogas plant affected the energy balance by only 4–6%. However, the energy ratio (GJ produced/GJ used) was decreased by 40–50%.
N balance and N loss
The N balance of Reference A represents an average of organic farms on sandy loam in Denmark in 2006. Since arable farms and dairy farms are combined, it is difficult to compare the results with other studies where dairy farms and arable farms are analyzed separately (e.g., Knudsen et al.Reference Knudsen, Kristensen, Berntsen, Petersen and Kristensen 40 ). For comparison, the annual farm N surplus of 47 organic farms in different countries ranged between 2 and 196 kg N ha−1 yr−1,Reference Watson, Bengtsson, Ebbesvik, Loes, Myrbeck, Salomon, Schroder and Stockdale 32 while our scenarios ranged from 79 to 106 kg N ha−1.
The nitrate-leaching loss was calculated as 30% of all N input to the field and is the same for all crops and crop rotations. It is well known that crop type and management play an important role for N-leaching losses. Askegaard et al.Reference Askegaard, Olesen, Rasmussen and Kristensen 41 found that leaching losses in an organic crop rotation depended on field management during autumn and winter. Smallest losses (on average 20 kg N ha−1) were observed when the soil was covered by cover crops during autumn and winter and highest losses were observed after stubble cultivation in autumn (on average 55 kg N ha−1). This indicates that management including cover crops would reduce N-leaching losses in all scenarios. In addition, use of cover crops could contribute to the biogas production and facilitate recycling of nutrients assimilated by the cover crop.
In general, the share of grass–clover in the crop rotations was high, 34–44%. Large amounts of N are often lost after plowing of grass–clover swards, especially if followed by winter wheatReference Eriksen, Askegaard and Kristensen 42 . Large areas of grass–clover in the crop rotations make management to reduce nitrate leaching more complicated. Since spring cereals are usually used for the establishment of grass–clover, there is a risk that grass–clover is to a greater extent followed by winter cereals. This is the case in S1. Therefore the crop rotations in S3 and S4 are more suitable for anaerobic digestion of grass–clover than S1, where large areas of grass–clover are already needed to supply grass–clover for the livestock.
GHG emissions
The largest contributions to the farm GHG emissions are CH4 emissions from enteric fermentation and N2O emissions from the field. Since the same feed intake is used in all scenarios, CH4 emissions from enteric fermentation depend exclusively on the number of animals on the farm. N2O emissions from the field depend of the amount of N added as slurry and digestate, and are therefore only affected by implementation of anaerobic digestion if lower amounts of total N are applied to the field due to the higher fertilizer value of the digestate. Another important factor for the N2O emission is pre-chain emissions connected with the imported feed, which is affected by the number of animals and the area for cash crop production on the farm. Therefore, apart from the reduced CO2 emissions, only emissions from the manure store are directly influenced by implementation of the anaerobic digester.
Field studies of the N2O emissions from cattle slurry compared with digested slurry are ambiguous. Carter et al.Reference Carter, Hauggaard-Nielsen, Heiske, Jensen, Thomsen, Schmidt, Johansen and Ambus 43 found emission factors at the same level for slurry and digested slurry, whereas Möller and StinnerReference Möller and Stinner 44 found higher emission from digested slurry than from raw cattle slurry. The calculated emissions of N2O from the field are based on emission factorsReference Olesen, Weiske, Asman, Weisbjerg, Djurhuus and Schelde 17 assuming 1% of added N is emitted as N2O. 35 In practice, N2O emissions are highly variable, and affected by, for example, tillage and soil typeReference Carter, Hauggaard-Nielsen, Heiske, Jensen, Thomsen, Schmidt, Johansen and Ambus 43 , Reference Chirinda, Carter, Albert, Ambus, Olesen, Porter and Petersen 45 . NO3 leaching is a large contributor to indirect N2O emissions from the field and reductions in NO3 leaching, as described in the section ‘N balance and N loss’, would therefore also reduce N2O emissions.
Loss of CH4 from the anaerobic digester affects the GHG balance of the system significantly. If the CH4 emission from the biogas plant increases from 1.8 to 2.8%, it will increase the GHG emission from the biogas plant (including manure stores) by 14.4% and the total GHG emission from the farm by 2.3%. In the current study, a loss of 1.8% is assumed, but other studies indicate that the loss might be higher: Kvist calculated a loss of 2.2% from the gas engine alone (personal communication, calculations based on results from Kvist et al.Reference Kvist, Frohn and Jørgensen 46 ). In addition, there is loss from the manure stores. Börjesson and BerglundReference Börjesson and Berglund 47 used a total loss of 1% of the produced CH4, but also estimated that if 10–20% of CH4 produced was lost, it would cause a GHG emission from the biogas system at the same level as the emission from fossil fuels. Thus, the CH4 loss from the biogas chain is an important factor to be aware of when environmental impacts of biogas systems are analyzed. Efforts to reduce the loss will significantly improve the biogas GHG balance.
Anaerobic digestion is connected to a loss of organic C as it is converted to CH4. In a study of the influence of biogas digestion on soil organic matter, MöllerReference Möller 48 found that digestion of slurry and crop residues reduced the overall soil C supply by approximately 33%. Both C incorporated directly into the soil as crop residues or cover crops and C added as slurry or digestate were taken into account. The amount of C added as slurry or digestate increased when slurry and crop residues were digested, but this was counteracted by a decrease in C incorporated directly into the soil. The amount of C applied was still enough to replace degradation of soil organic matter and maintain a positive C balance. In the current study, the amount of C added to the soil as slurry or digestate varied between 0.45 and 0.54 Mg ha−1 in the reference scenarios and 0.47–0.56 Mg ha−1 in the biogas scenarios (S2 and S4). This indicates that soil C changes due to change in applied C are not likely to happen. However, the risk of negative impact on soil N in S2 (Table 9) indicates a risk of a concurrent decline in soil C.
Conclusion
Implementation of biogas digestion changes the N flows on the farm by introducing new N sources from legume-based energy crops. The amount of NH4-N available for application on the farm cash crops increased when grass–clover was used for biogas production, but decreased when maize was used. A positive farm energy balance was obtained for all biogas scenarios, showing that biogas production on 10% of the area provides an energy surplus on the farm. Substantial reductions (up to 72%) in net GHG emissions were obtained as well. Biogas production on organic farms holds the possibility for the farms to achieve a positive energy balance and to provide self-sufficiency with organic fertilizer.
