| Front page | | Contents | | Previous | | Next |

Life Cycle Assessment of Slurry Management Technologies

Annex A. Reference Scenarios – Life Cycle Inventory data

• A.1 System Description and composition of reference slurry
  • A.1.1 System description
  • A.1.2 Composition of the reference slurry
  • A.1.3 Mass balances for N, P and K
  • A.1.4 Mass balances for Dry Matter, VS, Ash, Carbon, Cu and Zn
• A.2 In-house storage of slurry
  • A.2.1 System boundaries for the in-house storage of slurry
  • A.2.2 Emissions of CH4 and CO2
  • A.2.3 Emissions of NH3, N2O and other N compounds
  • A.2.4 Discharges to water and soil
  • A.2.5 Summary of the Life Cycle Inventory Data
• A.3 Storage
  • A.3.1 System boundaries and description of the process “Storage”
  • A.3.2 Emissions of CH4 and CO2
  • A.3.3 Emissions of NH3, N2O and other N compounds
  • A.3.4 Discharges to water and soil
  • A.3.5 Energy consumption for pumping and stirring
  • A.3.6 Electricity production
  • A.3.7 Summary of the Life Cycle Inventory Data
• A.4 Transport to field
  • A.4.1 System boundaries and description of the process “Transport to field”
  • A.4.2 Summary of the Life Cycle Inventory Data
• A.5 Field processes
  • A.5.1 System boundaries and description of the process “Field Processes”
  • A.5.2 Emissions of CH4 and CO2
  • A.5.3 Emissions of NH3, N2O and other N compounds
  • A.5.4 Emissions of N2O and NOX
  • A.5.5 Nitrogen leaching
  • A.5.6 Phosphorus leaching
  • A.5.7 Summary of the Life Cycle Inventory Data
• A.6 Avoided mineral fertilisers
  • A.6.1 Amount of replaced mineral fertilisers
  • A.6.2 Market considerations for mineral fertilisers
  • A.6.3 Marginal N fertiliser
  • A.6.4 Marginal P fertiliser
  • A.6.5 Marginal K fertiliser
  • A.6.6 Summary of the Life Cycle Inventory Data

A.1 System Description and composition of reference slurry

A.1.1 System description

The overall system for the reference scenario is described in section 3.1 of the report. The overall system description will not be repeated here. In addition to the system described in section 3.1, pumping and stirring have been included (i.e. pumping of slurry from the pre-tank to the outdoor storage, pumping from the outdoor storage to the transport tank before application to field and stirring of the slurry in the pre-tank and in the outdoor storage tank). See the flow diagram in figure A.1. The numbers in the figure refer to the numbers of the section in this Annex.

Figure A.1. Flow diagram for the reference scenario.

A.1.2 Composition of the reference slurry

The characteristics of the “reference slurry” for fattening pigs in the reference scenario are shown in table A.1 (per 1000 kg of slurry). The characteristics of the “reference slurry” for dairy cows are shown in table A.2 (per 1000 kg of slurry). The characteristics are given “ex animal”, “ex housing” and “ex storage”, see figure 3.2 in chapter 3. The references, assumptions and calculations are explained in the sections below.

Table A.1. Characteristics of slurry from fattening pigs in the reference scenario. Per 1000 kg of slurry “ex animal”, “ex housing” and “ex storage”.

^a Ash “ex storage” = 20% of DM

^b VS “ex storage” = 80% of DM

^c For pig slurry, the content of NH4⁺-N ex storage corresponds to 75% of the total N content (Poulsen et al. (2001) and DJF (2008b))

Table A.2. Characteristics of slurry from dairy cows in the reference scenario. Per 1000 kg of slurry “ex animal”, “ex housing” and “ex storage”.

^a Ash “ex storage” = 20% of DM

^b VS “ex storage” = 80% of DM

^c For cattle slurry, the content of NH4⁺-N ex storage corresponds to 60% of the total N content (Poulsen et al. (2001) and DJF (2008b))

The data for dry matter (DM), nitrogen (N), phosphorus (P) and potassium (K) are based on the Danish Normative system for assessing manure composition (Poulsen et al. (2001), DJF (2008a) and DJF (2008b)). It is however acknowledge that these values might differ notably from the composition of pig and cattle slurry measurements “in the real world” due to differences in e.g. diets and slurry handling. Therefore, the “ex storage” values from the Danish Normative system have been compared to measurements from Knudsen and Birkmose (2005) and Hansen et al. (2008), as shown in table A.3. The “ex storage” values are the values compared since this is what was available in the literature for comparison. The measurements by Knudsen and Birkmose (2005) are based on 55 samples of pig slurry and 50 samples of cattle slurry. According to Birkmose (2008, personal communication), the pig slurry samples are from all kind of pig farms (including mixed farms with sows, piglets and fattening pigs). The cattle samples are primarily based on dairy farms, but for these, calves are often included. In their study, the storage time and method for the slurry varies. As it can be seen from table A.3, there are considerable variations on the minimum and maximum for the measurements. The data in Hansen et al. (2008) are based on more than 270 samples of pig slurry and 200 samples of cattle slurry. The measurements were made right before the application to field (Personal communication, Hansen, 2009).

Table A.3. Values from DJF (2008a) compared to measurements (Knudsen and Birkmose, 2005) and Hansen et al. (2008) for selected characteristics of pig and cattle slurry. Uncertainty range appears in brackets [ ]. All data are given in kg per 1000 kg slurry ex storage.

From table A.3, it can be seen that the concentrations for fattening pig slurry from the Danish Normative system in general is higher than the “average pig farms” measured by Knudsen and Birkmose (2005) and Hansen et al. (2008). This is probably due to the fact that Knudsen and Birkmose (2005) includes all kind of pig farms and mixed farms including sows and piglets and, again, differences in feeding, housing systems, slurry handling, slurry storage time etc. Furthermore, the Danish Normative system does not include water added in the housing systems, as described below. Thus, the concentrations “ex storage”are higher as the slurry is less diluted.

The dairy cow slurry from the Danish Normative system has a significant higher content of dry matter and a higher concentration of nitrogen (N), phosphorous (P) and potassium (K) than the measurements made by Knudsen and Birkmose (2005) and Hansen et al. (2008). Actually, all the DJF (2008) values for dairy cows are higher than the upper limit of the uncertainty range provided by Knudsen and Birkmose (2005). Knudsen and Birkmose (2005) suggest that the difference between the measured data and the data from the Danish Normative system is probably due to the fact that more water is actually added (in the housing systems or during outdoor storage) than what is included in the Norm data. Poulsen et al. (2001) do not include added and lost water in the housing systems in their calculations. They estimate the values in tables, but do not include these values in the calculations. The same problem appears in the data from DJF (2008a). Poulsen et al. (2001, page 96) state that due to lack of data, the loss of water and evaporation of water is not included in the calculations and instead of calculating dry matter it is set in accordance with measurements from “real life”.

It has not been possible to perform sensitivity analysis for variations of slurry compositions, as the slurry composition influences on all mass balances and all emissions throughout the report and in all Annexes. This would require a modelling tool, which has not been established within the frames of the budget.

There are also unexplained deviations in the mass balances in the DJF (2008a) data for the outdoor storage (added rain) (see table A.4 below). Jacobsen et al. (2002) state that the method used by Poulsen et al. (2001) is problematic as the amount of slurry (”ex animal”) and inputs and outputs from the housing system and during the storage do not necessarily yield the amount “ex storage” that is given in the Norm Data. The problem appears for all years, also data in DJF (2008a). Jacobsen et al. (2002, table 2.5) calculated the mass balances and found deviations ranging between -77% to +32%.

In table A.4 the mass balances used in DJF (2008a) are shown. Furthermore, mass balances when including water added in the housing system are shown. The amount of added water is based on data from Poulsen et al. (2001). However, the mass balances for including water in the housing systems give too high amount of water as there is no data for the evaporation.

When regarding table A.4 it seems that the mass balances by DJF (2008a) include inadequate amounts of water, especially for dairy cow slurry, and it is likely that this is part of the reason for the difference between the Norm Data and the measurements as shown in table A.3.

As described in chapter 3, it is assumed that the slurry tank for outdoor storage is covered by a floating layer of straw (pig slurry) or a natural floating layer (cattle slurry). These covers do not prevent rain from diluting the slurry. According to Poulsen et al. (2001, page 128) the amount corresponds to approximately 110 litres of water per 1000 kg slurry “ex housing”. However, when using the Norm Data for 2008 (DJF, 2008), the amount of water applied is significantly lower for dairy cows (4.4% as can be seen from table A.4).

It spite of this, it has been chosen to use the Norm Data without corrections for the water amounts in this study, as the Norm Data are “Danish standard data” used for the majority of the Danish slurry studies and as it is not within the frames of this report to improve the Danish Norm Data nor to claim a better knowledge of slurry composition.

Table A.4. Mass balances for the total volume of slurry “ex animal”, “ex housing” and “ex storage” used in this study and from DJF (2008a).

^a The data in DJF (2008a) is rounded off (e.g. 0.47 instead of 0.474). When using the rounded data for the calculations, the “ex storage” values are not exactly the same as in DJF (2008a). The dilution factor that fits best with the “ex storage” data is that 8.6% water is added.

^b For fattening pigs (at fully slatted floors), the amount of wasted drinking water is estimated to 75 liters per pig and cleaning water to 30 litres per pig (DJF (2008b) and Poulsen et al. (2001)). As one pig produces 0.47 tonnes of slurry (DJF (2008a)), the water wasted in the housing system corresponds to 105 litres/0.47 tonnes = 223 kg per 1000 kg.

^c The net amount of rain (i.e. rain minus evaporation) is 110 kg per 1000 kg slurry “ex housing” (Poulsen et al. (2001, page 128-130) and DJF (2008b).

^d For dairy cows, slurry based housing unit, the amount of wasted drinking water is estimated to 100 litres per cow per year and cleaning water to 3000 litres per cow per year (DJF, 2008b, and Poulsen et al., 2001). As one cow produces 20.4 tonnes of slurry per year (DJF, 2008a), the water wasted in the housing system corresponds to 3100 litres/20.4 tonnes = 152 kg per 1000 kg slurry.

A.1.3 Mass balances for N, P and K

In the Danish Normative system for assessing manure composition, the data for nitrogen (N), phosphorus (P) and potassium (K) are given “ex animal” and “ex storage”. This study is based on the “ex animal” and “ex storage” values from DJF (2008a). The “ex housing” values are calculated by establishing mass balances from “ex animal” to “ex housing system” and from this to “ex storage” following the natural flow of the slurry:

The calculations for N, P and K are shown in table A.5 for fattening pig slurry and in table A.6 for dairy cow slurry. The explanations are given in the text and in the notes for the tables.

The Total-N in the slurry “ex animal” and “ex storage” is based on data from DJF (2008a).

The total-N in the slurry “ex housing” is calculated in accordance with the preconditions in Poulsen et al. (2001). For fattening pigs in housing units with fully slatted floors, the in-housing NH₃ emission corresponds to 16% (NH₃-N in pct. of the total N ex animal). For dairy cows in cubicle housing systems with slatted floor (1.2 m channel) the emission factor is 8%(NH₃-N in pct. of the total N ex animal). The NH₃-emissions during storage of slurry correspond to 2% (NH₃-N in pct. of the total-N ex housing) for both pig slurry and cattle slurry. In DJF (2008b), new preconditions are given for the NH₃ emissions, however, when calculating the N balances for fattening pigs and dairy cows in DJF (2008a), it seems that the new preconditions from DJF (2008b) have not been used. Accordingly, the preconditions from Poulsen et al. (2001) have been used in this report.

As mentioned in section 3.2, the NH₄⁺-N in the slurry “ex housing” and “ex animal” has not been estimated in this report, as data on this has not been identified and as balances on NH₄⁺ could not be established. As there are microbial metabolisms and biochemical processes transforming organic N to inorganic N it is not reasonable to assume that the relative amount of NH₄⁺-N “ex housing” and “ex animal” is the same as “ex storage”.

In the housing units for dairy cows, straw is added as bedding material. According to Poulsen et al. (2001) the amount is 1.2 kg per animal per day for slurry based housing units for dairy cows. Adding of straw affects the mass balances for nitrogen, phosphorous, and potassium, see table A.6.

According to Poulsen et al. (2001) and DJF (2008b), straw is not added to fattening pigs in fully slatted floor housing systems.

For pig slurry, straw is added as a floating layer during storage in order to reduce the emissions, as described in chapter 3. It is assumed that straw is not added to slurry tanks for dairy cows. Cut straw is added as floating layer to cover pig slurry during storage corresponds to 10 kg straw per m² slurry surface (Rasmussen et al., 2001). With a 4 m deep slurry tank this corresponds to 2.5 kg straw per 1000 kg slurry. The added straw increases the total mass slightly (2.5 kg per 1000 kg). The added mass of the straw is less than 0.3% of the total mass ¹. This is insignificant for the overall results and the inclusion would reduce the transparency of the calculations more than it would increase the precision of the results. The amounts added by the straw are insignificant compared to the differences between the compositions of slurry from farm to farm. For dry matter, the added amount corresponds to 3.4% ², which is regarded as insignificant compared to the overall uncertainty.

The amount of N, P and K added in the straw for the floating layer for outdoor storage of pig slurry is insignificant. ³ Accordingly, the mass, dry matter, nitrogen, phosphorus and potassium added by the straw during outdoor storage have been ignored in this study. Added straw during outdoor storage is not included in the mass balances in DJF (2008) either.

Phosphor (Total-P) and potassium (K) “ex animal” and “ex storage” are based on DJF (2008a). For dairy cows slurry, P and K is added by the added straw in the housing system. The amount of Total P and K is assumed to be unchanged during outdoor storage.

Table A.5. Calculation of the “ex housing” characteristics of slurry from fattening pigs (reference scenario)

^a See table A.4

^b N ex animal: 3.10 kg / 0.47 tonnes = 6.60 kg N per 1000 kg slurry (DJF, 2008a).

^c The in-housing NH3-emissions is 16% (pct. NH3-N of total-N ex animal) (Poulsen et al., 2001) = 6.60 kg * 0.16 = 1.06 kg

^d NH3-emissions during storage is 2% (pct. NH3-N of total-N ex housing) (Poulsen et al., 2001) = 5.54 * 0.02 = 0.11 kg

^e P ex animal: 0.53 kg / 0.47 tonnes = 1.13 kg P per 1000 kg slurry (DJF, 2008a).

^f K ex animal: 1.34 kg / 0.47 tonnes = 2.85 kg K per 1000 kg slurry (DJF, 2008a).

^g The calculated potassium value (2.85 kg/1.086 = 2.62 kg) does not fit exactly to the DJF (2008a) “ex storage” value (2.60 kg). As the difference is less than 1%, the DJF (2008a) value is used.

i In-house emissions, see table A.9. NH3-N emissions: 1.06 kg. N2O-N: 0.013 kg. NO-N: 0.013 kg. N2-N: 0.039 kg. In total: -1.125 kg

j Storage emissions, see table A.11. NH3-N emissions: 0.11 kg. N2O-N: 0.033 kg. NO-N: 0.033 kg. N2-N: 0.099 kg. In total: -0.275 kg

Table A.6. Calculation of the “ex housing” characteristics of slurry from dairy cows (reference scenario)

^a See table A.4

^b N ex animal: 140.2 kg / 20.4 tonnes = 6.87 kg N per 1000 kg slurry (DJF, 2008a).

^c The in-housing NH3-emissions is 8% (pct. NH3-N of total-N ex animal) (Poulsen et al., 2001) = 6.87 kg * 0.08 = 0.55 kg.

^d According to Poulsen et al (2001), there is a consumption of 1.2 kg straw per cow per day for bedding. Straw contains 85% dry matter and 0.005 kg N per kg dry matter (Poulsen et al., (2001). Accordingly, the added amount of N during a year for a dairy cow is: 1.2 kg straw per animal * 365 days * 0.85 kg DM/kg straw * 0.005 kg N/kg DM = 1.8615 kg N. This amount corresponds to the total amount from a dairy cow during a year. As the dairy cow produces 20400 kg slurry (DJF (2008a), the amount corresponds to 0.09 kg N per 1000 kg slurry. Note that according to DJF (2008b) 0.4 kg straw is added per day per animal, but it seems like DJF (2008a) has used the 1.2 kg straw per day from Poulsen et al. (2001).

^e NH3-emissions during storage is 2% (pct. NH3-N of total-N ex housing) (Poulsen et al., 2001) = 6.41 * 0.02 = 0.13 kg

^f P ex animal: 20.8 kg / 20.4 tonnes = 1.02 kg P per 1000 kg slurry (DJF, 2008a).

^g K ex animal: 118.6 kg / 20.4 tonnes = 5.81 kg K per 1000 kg slurry (DJF, 2008a).

^h The P content in straw is 0.00068 kg P per kg dry matter. The calculations are parallel to the calculations for N above: 1.2 kg straw per animal * 365 days * 0.85 kg DM/kg straw * 0.00068 kg P/kg DM / 20400 kg slurry = 0.012 kg P per 1000 kg slurry.

ⁱ The K content in straw is 0.01475 kg K per kg dry matter. The calculations are parallel to the calculations for N above, however, it seems like that DJF (2008a) has used the amount of straw from DJF (2008b) i.e. 0.4 kg straw instead of the amount from Poulsen et al. (2001). 0.4 kg straw per animal * 365 days * 0.85 kg DM/kg straw * 0.01475 kg K per kg DM / 20400 kg slurry = 0.09 kg K per 1000 kg slurry.

j In-house emissions, see table A.9. NH3-N emissions: 0.55 kg. N2O-N: 0.014 kg. NO-N: 0.014 kg. N2-N: 0.042 kg. In total: -0.53 kg

k In-house emissions, see table A.11. NH3-N emissions: 0.13 kg. N2O-N: 0.034 kg. NO-N: 0.034 kg. N2-N: 0.10 kg. In total: -0.3 kg

A.1.4 Mass balances for Dry Matter, VS, Ash, Carbon, Cu and Zn

For dry matter, Poulsen et al. (2001) and DJF (2008a) only give data ”ex storage”. This also applies for all the data from other references, i.e. data are given “ex storage”. Accordingly, the “ex housing” values and “ex animal” values have to be calculated “backwards”:

The calculations for the rest of the parameters are shown in table A.7 (pigs) and A.8 (cows). The assumptions are described in the following text and the calculations are shown in the footnotes to the tables and to the text.

The dry matter (DM) content “ex storage” is based on the Danish Normative system for assessing manure composition (DJF, 2008a).

Dry matter “ex housing” and “ex animal” is calculated in accordance with the losses based on mass balances:

Table A.7. Calculation of the “ex housing” and “ex animal” characteristics of slurry from fattening pigs for the reference scenario, for selected components other than N, P or K.

^a Adjusted by the relative amount of slurry: “amount of slurry ex storage”/“amount of slurry ex housing”

^b Adjusted by the relative amount of slurry: “amount of slurry ex storage”/“amount of slurry ex animal”

^c DM loss during storage corresponds to 5% of the “ex housing” value = 61 kg * 0.05/(1-0.05) = 3.2 kg

^d In-housing DM loss corresponds to 10% of the “ex animal” value = 64.2 kg * 0.10/(1-0.10) = 7.1 kg

^e It is assumed that the loss of volatile solids is identical to the loss of Dry Matter

^f It is assumed that all the volatile solids lost are easily degradable volatile solids

^g For pig slurry ex animal, 65% of the VS is easily degradable and 35% is heavily degradable, see text below.

^h Assumption for pig slurry: 47.9% of dry matter is C, see text below.

ⁱ Carbon loss during storage is assumed to be in the same order as the DM loss, i.e. 5% of the “ex housing” value. C loss during storage = 29.2 kg * 0.05/(1-0.05) = 1.5 kg

^j The in-housing carbon loss is assumed to be in the same order as the DM loss, i.e. 10% of the “ex animal” value. C loss during storage = 30.7 kg * 0.10/(1-0.10) = 3.4 kg

k For pig slurry 0.0453 % of the dry matter is copper and 0.135 % is zinc, see text below.

Table A.8. Calculation of the “ex housing” and “ex animal” characteristics of slurry from dairy cows for the reference scenario, for selected components other than N, P or K..

^a Adjusted by the relative amount of slurry: “amount of slurry ex storage”/“amount of slurry ex housing”

^b Adjusted by the relative amount of slurry: “amount of slurry ex storage”/“amount of slurry ex animal”

^c DM loss during storage corresponds to 5% of the “ex housing” value = 103 kg * 0.05/(1-0.05) = 5.4 kg

^d In-housing DM loss corresponds to 10% of the “ex animal” value = 108.4 kg * 0.10/(1-0.10) = 12.0 kg

^e It is assumed that the loss of volatile solids is identical to the loss of dry matter

^f It is assumed that all the loss of volatile solids is easily degradable volatile solids

^g For cattle slurry ex animal, 48% of the VS is easily degradable and 52% is heavily degradable, see text below.

^h Assumption for cattle slurry: 43.9% of dry matter is C, see text below.

ⁱ Carbon loss during storage is assumed to be in the same order as the DM loss, i.e. 5% of the “ex housing” value. C loss during storage = 45.2 kg * 0.05/(1-0.05) = 2.4 kg

^j The in-housing carbon loss is assumed to be in the same order as the DM loss, i.e. 10% of the “ex animal” value. C loss during storage = 47.6 kg * 0.10/(1-0.10) = 5.3 kg

k For cattle slurry, 0.0113 % of the dry matter is copper and 0.0217 % is zinc.

During storage of slurry in-house and outdoor, there is a loss of dry matter due to microbial metabolisms and biochemical degradation. Poulsen et al. (2001) (page 130) and DJF (2008b) use an estimate for the in-housing loss of dry matter at 10% of the “ex animal” content of dry matter for slurry from pigs and cattle. The loss during storage is estimated to 5% of the “ex housing” content of dry matter for slurry (for pigs and cattle). The magnitudes of the losses depend to a great degree on the residence time for the slurry and the temperature. There are probably relatively high uncertainties related to this estimate, however, it has not been possible to identify better data for the loss of dry matter.

The ash is assumed to constitute 20% of the dry matter “ex storage” (estimate by Sommer et al. (2008)). It is assumed that there is no change in the ash amount from “ex animal” to “ex storage”.

The volatile solids (VS) are calculated as 80% of the dry matter “ex storage”. This is a rough estimate (Sommer et al., 2008) - but is nonetheless well in accordance with measurements made by S O Petersen (Personal communication with S O Petersen, December 2008).

The share of easily degradable VS and heavy degradable VS for pig slurry is based on Sommer et al. (2001, Appendix 5). Sommer et al. (2001) estimate that for pig slurry, 65% of the volatile solids are easily degradable and 35% are heavy degradable (”ex animal” values). For cattle slurry, 48% of the volatile solids are easily degradable and 52% are heavy degradable (”ex animal” values). Sommer et al. (2001) assume that all the loss of dry matter corresponds to loss of easily degradable VS.

Thus, the share of easily degradable VS and heavy degradable VS “ex housing” and “ex storage” can be calculated for fattening pigs and for dairy cows ⁴.

Data for carbon might be based directly on the measurements by Knudsen and Birkmose (2005): 28.1 kg C per 1000 kg cattle slurry and 18.2 kg C per 1000 kg pig slurry. However, the content of dry matter measured by Knudsen and Birkmose is significantly lower than the norm data by DJF (2007). Accordingly, it has been assumed that the ratio between carbon and dry matter from Knudsen and Birkmose (2005) can be used i.e. for pig slurry 47.9% of the dry matter is carbon and for cattle slurry, 43.9% of the dry matter is carbon ⁵. It is assumed, that the loss of carbon is in the same proportion as the loss of dry matter, i.e. 5% during storage (DJF, 2008) and 10% in the housing units.

Data for copper and zinc are based on Knudsen and Birkmose (2005). The data from Knudsen and Birkmose (2005) are at the same level as Møller et al. (2007) (zink in pig slurry is twice the amount in Knudsen and Birkmose (2005), which is still regarded as “at the same level” due to the high variations for the slurry). As argued above, the slurry defined by the Normative System (DJF, 2008a and Poulsen et al. 2000) is probably more concentrated than the measured data by Knudsen and Birkmose (2005) due to the fact that water in the housing units is not included in the Normative System. As for carbon, it is assumed that the ratio between copper and dry matter and the ratio between zinc and dry matter from Knudsen and Birkmose (2005) can be used, i.e.: for pig slurry 0.0453 % of the dry matter is copper and 0.135 % is zinc and for cattle slurry, 0.0113 % of the dry matter is copper and 0.0217 % is zinc. It is assumed that there is no gain or loss during storage.

The pH is set to 7.8 for the reference system for both pig slurry and cattle slurry, based on the measurements by Sommer and Husted (1995). Sommer and Husted (1995) found an average pH of 7.75 [7.2-8.3] for pig slurry and average pH of 7.84 [7.7-8.1] for cattle slurry. The raw pig slurry and cattle slurry was collected in the channels below the slatted floor of the animal housing.

The density of slurry has been set to 1053 kg per m³, based on the study by Sherlock et al. (2002). Lopez-Ridaura et al. (2008) use a density of 1034 kg per m³, which is not far from the value given by Sherlock et al. (2002).

The density will most probably depend on a lot of factors (e.g. water, organic matter and salts in the slurry as well as it is affected by diets and management).

A.2 In-house storage of slurry

A.2.1 System boundaries for the in-house storage of slurry

The life cycle inventory data for the slurry in the housing units includes the emissions from the slurry only (not including enteric fermentation) in accordance with the system boundaries described in chapter 2. The Life Cycle Inventory data for the in-house storage of slurry are shown in table A.9.

A.2.2 Emissions of CH₄ and CO₂

The CH₄ emissions from slurry in the housing units are based on the IPCC (2006) Tier 2 approach. According to this the CH₄ emissions are 3.29 kg CH₄ per 1000 kg pig slurry and 2.85 kg CH₄ per 1000 kg dairy cow slurry ⁶

The uncertainty on the CH₄ emissions is high. The IPCC (2006) model use a very rough partitioning in “storage < 1 month” and “storage > 1 month”.

The emission factors (in kg CH₄ per kg VS) is 5.67 times higher for “storage > 1 month” than for “storage > 1 month” which is an unrealistic jump. The emission of CH₄ should rather have been modelled as a function of time. In this study, the high emission factor from IPCC (2006) has been used as a conservative estimate. The CH₄ emissions depend on a range of factors, among these the CH₄ emissions to a great extent depend on the retention time in the housing units and on the biological activity. However, in the present study, the CH₄ emissions for the slurry management technologies are calculated relatively to the CH₄ emissions in the reference scenario, the significance of the uncertainty is therefore reduced slightly for the comparisons to the new technologies. Sensitivity analysis is carried out for the CH₄ emissions.

The emissions of CO₂ are based on a very rough estimate, as no data have been found. Sommer et al. (2008) state: “In most inventories or scenario calculations of carbon loss due to gaseous emissions are expressed as DM loss. The justification is partly that carbon is a major constituent of the organic matter in the DM fraction. Slurry is as mentioned an anaerobic matrix in which organic carbon transformation is relatively slow. The absence of oxygen is a precondition for the production of CO₂ and CH₄ via microbial metabolisms of organic material in livestock and livestock manure. During slurry storage inside the animal houses reduction in organic components will be affected by slurry removal frequency i.e. storage time and temperature. There are few studies of the reduction of DM under realistic conditions so in the Danish Normative system for assessing manure composition it has been decided that a rough estimate of DM loss from slurry stored in house is 10% (Poulsen et al. 2001).”

The CO₂ emissions are calculated as the total loss of carbon in the housing units minus the carbon lost as CH₄ emissions. The total carbon loss is calculated in table A.7 and A.8. The CO₂ emissions from the slurry in the slurry pits are estimated to 3.44 kg CO₂ for pig slurry ⁷ and 11.6 kg CO₂ for dairy cow slurry ⁸.

The uncertainty on the CO₂ emissions is very high, however, as methane has a much higher greenhouse gas potential (approximately 23 times as high as CO₂) , the uncertainty on the CO₂ is not very important as the CH₄ emissions from the process is the main contributor to the global warming impact.

A.2.3 Emissions of NH₃, N₂O and other N compounds

The emissions of NH₃ are based on data from the Danish Normative system for assessing manure composition (Poulsen et al. (2007) and DJF (2007)). According to Poulsen et al. (2001), the emission factor for fattening pigs (fully slatted floor) is 16% NH₃-N of the total-N “ex animal”. Thus, the emission of NH₃ from the slurry in the housing system is 1.06 kg NH₃-N per 1000 kg pig slurry “ex animal” (see calculations in table A.5). This value is at the same level as measured values by Kai et al. (2008) at 0.91 g NH₃-N per 1000 kg pig slurry “ex animal” ⁹. It is, however, slightly lower than the emission coefficients suggested by Sommer et al. (2006), which corresponds to an emission of 1.3 NH₃-N per 1000 kg pig slurry “ex animal” ¹⁰.

The NH₃ emission factor for dairy cows (Cubicle housing system with slatted floor, 1.2 m channel)” is 8% NH₃-N of the total-N (excretion) (Poulsen et al., 2001).Thus the emission of NH₃ from the slurry in the housing system is 0.55 kg NH₃-N per 1000 kg dairy cow slurry “ex animal” (see calculations in table A.6). Sommer et al. (2006) suggest an emission coefficient for cattle in cubicle housing units with slatted floors, corresponding to an emission of 0.68 NH₃-N per 1000 kg dairy cow slurry “ex animal ” ¹¹, which is only 23% higher than the 0.55 kg NH₃-N mentioned above.

The direct N₂O emissions in this study are based on IPCC (2006, table 10.21). IPCC (2006) estimates the N₂O emissions from pit storage below animal confinements to be 0.002 kg N₂O-N per kg N “ex animal” (uncertainty: a factor 2), based on the judgement of an IPCC expert group combined with various studies.

The IPCC (2006) recommend including the indirect N₂O emissions, see description in IPCC (2006)¹². The indirect N₂O emission corresponds to 0.01 kg N₂O–N per kg (NH₃–N + NO_X–N volatilised) (IPCC, 2006, table 11.3).

Dämmgen and Hutchings (2008) developed a new approach for assessing emissions of gaseous nitrogen species from manure management. In their study, they assume that the emission of nitrogen monoxide (NO) is at the same level as the direct emissions of nitrous oxide (N₂O) (measured as NO-N and N₂O-N). Furthermore, they assume that emission of nitrogen (N₂) is three times as high as the direct emissions of nitrous oxide (N₂O) (measured as N₂-N and N₂O-N). In the current study, it has not been possible to find data on NO₂, however, due to the considerable uncertainty on the estimates on the NO emissions, it is assumed that the NO emissions represent the total NO_X emissions (NO_X = NO + NO₂).

A.2.4 Discharges to water and soil

For both pig slurry and dairy cow slurry, it is assumed, that there are no emissions to water and soil from housing systems in the reference scenario, as leakages from housing systems are prohibited in Denmark (Poulsen et al. (2001), page 117).

A.2.5 Summary of the Life Cycle Inventory Data

The Life Cycle Inventory Data for storage of the slurry in the housing units are shown in table A.9. Feed for the animals, medicine, straw for bedding and water consumption are not included within the system boundary of the reference scenario.

Table A.9. Life cycle Inventory data for storage of slurry in the housing units (reference scenario). All data per 1000 kg of slurry “ex animal”

A.3 Storage

A.3.1 System boundaries and description of the process “Storage”

The process called “Storage” includes emissions from:

• Storing slurry in the pre-tank (typically 2-6 weeks).
• Storing slurry in the outdoor storage for months before application to fields.

Furthermore, the energy consumption from pumping and stirring is included, i.e.:

• Flushing slurry from the slurry pits in the housing units to a pre-tank.
• Stirring slurry in the pre-tank before pumping to the outdoor storage.
• Pumping slurry from the pre-tank to the outdoor storage by a pump.
• Stirring slurry in the outdoor concrete tank when straw is added (pig slurry only)
• Stirring slurry before pumping from outdoor storage tank.
• Pumping slurry from the storage tank to the transport tank.

The emissions from these storage processes are treated together, as the available literature data for emissions are joined under “emissions during storage”. It would have been ideal to separate the pre-tank emissions from the outdoor storage emissions, but unfortunately, it has not been possible.

It is assumed that the emissions during this handling is negligible compared to the emissions from the outdoor storage.

The materials for the pre-tank (concrete etc.) are not included, as it is assumed to be more or less the same for in the reference scenario and for the alternative technologies. Preliminary calculations have shown that the construction of a concrete pre-tank (divided by the slurry amounts passing through the pre-tank during the life time of the pre-tank) is insignificant for the overall environmental impacts for slurry management.

In the reference scenarios, it is assumed that the slurry is stored in an outdoor concrete tank. LCA data for the concrete slurry storage tank is based on the Ecoinvent process: “Slurry store and processing, operation” (300 m³ concrete vessel, average life time 40 years). The process includes the production of a concrete vessel (divided by the amounts of slurry it contains during 40 years of use), production of a screw agitator and the electricity for stirring (the slurry is normally stirred before application to fields). The Ecoinvent data for the concrete store actually encompass a covered, under-floor slurry store, however, as preliminary calculations in SimaPro 7.1 has shown, that the slurry store is of minor significance for the overall results, it is acceptable.

As described in chapter 3, it is assumed that pig slurry is covered by cut straw. The amount corresponds to 2.5 kg per 1000 kg pig slurry (see section A.2). As straw is regarded as a waste product from grain production (rather than a co-product) the life cycle data of straw production is not included.

The energy consumption for cutting and adding straw has been left out as it is regarded as insignificant (there are only 2.5 kg straw per 1000 kg slurry i.e. less than 1% of the weight). As mentioned before, it is assumed that it is not necessary to add cut straw to the cattle slurry (Rasmussen et al., 2001, page 31).

No additional chemicals or additives are added. During the storage, rain is adding water to the slurry. Accordingly, the total amount of slurry is slightly higher after storing, as described in the mass balances chapter 3. There are no wastes or by-products from the process.

A.3.2 Emissions of CH₄ and CO₂

The CH₄ emissions from outdoor storage of slurry are based on the IPCC (2006) Tier 2 approach. According to this the CH₄ emissions are 1.94 kg CH₄ per 1000 kg pig slurry and 1.68 kg CH₄ per 1000 kg dairy cow slurry ¹³

The CH₄ emissions based on IPCC (2006) are at the same level as modelled by Sommer et al. (2001). For the outdoor storage of slurry, Sommer et al. (2001) modelled the CH₄ emissions to 2.07 kg CH₄ per 1000 kg pig slurry ¹⁴ and 1.61 kg CH₄ per 1000 kg cattle slurry¹⁵. There is significant uncertainty related to the magnitude of the CH₄ emissions as discussed in Olesen et al. (2004, Annex B). A scientific discussion regarding various estimates and investigations is beyond the scope of this study.

The liquid slurry is usually homogenized (stirred) in the tank prior to application. Mixing may release H₂S and CH₄. It is assumed that the amount of CH₄ emissions during stirring are minor compared to the loss during storage, hence, only emissions during storage and application is included in this study.

The CO₂ emissions for the outdoor storage are calculated as the in-housing CO₂ emissions in section A.2, i.e. the total loss of carbon in the housing units minus the carbon lost as CH₄ emissions. The total carbon loss is calculated in table A.7 and A.8. The CO₂ emissions from the slurry in the slurry pits are estimated to 0.18 kg CO₂ for pig slurry ¹⁶ and 4.21 kg CO₂ for dairy cow slurry ¹⁷.

Loyon et al. (2007) measured gaseous emissions from aerobic treatment of pig slurry and compared this to the emissions from a conventional storage system. Calculations on their results from the “conventional storage system” show that the CO₂-C emissions corresponds to78% of the CH₄-C emissions (i.e. when there are 100 grams of CH₄-C emissions there will be 78 grams of CO₂-C). Sneath et al. (2006) monitored green house gas emissions from covered manure stores on dairy farms. When comparing the CO₂ and CH₄ emissions from Sneath et al. (2006) it can be seen that for each time, 100 grams of CH₄-C is emitted, approximately 120 grams of CO₂-C is emitted. The ratio of CH₄ and CO₂ depends on the proportion of the biological processes. Møller et al. (2004) found a high biological degradation in the aerobic surface layers of the stored manure at 15°C leading to CO₂ emissions. As mentioned in section A.3, the uncertainty on the CO₂ emissions is very high however, as methane is a greenhouse gas, with a much higher global warming potential as compared with CO₂, the uncertainty on the CO₂ is not very important, since the CH₄ emissions from the process are in the main contributor to the global warming.

A.3.3 Emissions of NH₃, N₂O and other N compounds

The NH₃ emissions are based on Poulsen et al. (2001, page 119). Poulsen et al. (2001) presume that the slurry tank is 4 meter deep and that the storage time is 12 months (page 128 in Poulsen et al, 2001). A storage time of 12 months might be “in the high end”. According to Poulsen et al. (2001), the emission of NH₃–N is 2% of the total-N in the slurry “ex housing” for both pigs and cattle (see the calculations in table A.5 and A.6).

The direct N₂O, emissions are based on IPCC (2006). IPCC (2006) recommend an emission factor of 0.005 kg N₂O-N per kg N “ex animal” for slurry stored with natural crust cover. IPCC (2006) estimate the uncertainty to be a factor 2. As the IPCC factor is “ex animal”, it means that difference in the various housing units and the biological degradation in the housing units (which might change the total content of N in the slurry) is not taken into consideration. However, it has not been possible to make a better approximation within the framework of this study.

In addition, the indirect N₂O emission has been included in accordance with the IPCC (2006) guidelines, as described in section A.2, i.e. 0.01 kg N₂O–N per kg (NH₃–N + NO_X–N volatilised) (IPCC, 2006, table 11.3).

The NO and N₂ emissions are based on the rough estimate by Dämmgen and Hutchings (2008), as for the NO and N₂ emissions in the housing units, see section A.3.

A.3.4 Discharges to water and soil

It is assumed, that there are no emissions to water and soil from slurry storage in the reference scenario, as leakages from slurry tanks are prohibited in Denmark (Poulsen et al. (2001), page 117).

A.3.5 Energy consumption for pumping and stirring

The energy consumption for pumping and stirring in the reference scenario is shown in table A.10.

The energy and water for flushing the slurry from the slurry pits in housing units to the pre-tank is not included. It is assumed that it is more or less identical in the reference scenario and in the scenarios for the alternative technologies and that a potential difference in how this is done is insignificant for the overall environmental impacts for slurry management.

It is assumed that the energy consumption for the stirring is 1.22 kWh [0.71-2.41] per 1000 kg slurry (Personal communication with J Mertz (2008) based on communication with farmer). The Ecoinvent database contains the process “Slurry store and processing”, which contains data for a covered, under-floor slurry store including a 6 kW marine screw agitator. According to the Ecoinvent data, the energy consumption used by the agitator is approximately 0.4 kWh per 1000 kg slurry. In this study, the energy consumption for stirring is assumed to be 1.2 kWh per 1000 kg slurry.

According to Sandars et al. (2003), the energy consumption for pumping slurry in a pipeline from housing to storage is in the range of 0.2-0.5 kWh per 1000 kg slurry. They base their data on various pump manufacturers. Obviously, the energy required depends on factors like slurry density, distance travelled, flow rate, velocity etc. The transport from the slurry tank to the slurry transport tanker can alternatively be carried out by the use of a tractor with a diesel engine corresponding to a consumption of 60-70 litres for 1000 m³ slurry (which corresponds to approximately 0.5 kWh per 1000 kg slurry) ¹⁸. In this study, the energy consumption for pumping is presumed to be 0.5 kWh electricity per 1000 kg slurry.

Data for producing the pump has not been included. Preliminary calculations in SimaPro 7.1 by the use of the Ecoinvent data mentioned above (for the process “Slurry store and processing”) showed that the production of the pump was insignificant compared to the energy consumption during use of the pump.

When adding straw to pig slurry for floating layer, stirring is required (by law) and accordingly, stirring is included twice for pig slurry in the storage tank (i.e. when adding straw and before pumping the slurry to the transport container). The total energy consumption for stirring and pumping is shown in table A.10.

Table A.10 Energy consumption for stirring and pumping slurry during storage. All data are expressed per 1000 kg of slurry “ex housing”.

A.3.6 Electricity production

The modelling of marginal electricity in Denmark is based on Lund et al. (2009). According to this, the marginal electricity shall be modelled as be modelled as “Business as Usual + Power Plant Natural gas” (table 3 in Lund et al. ,2009), i.e. 1% wind, 49% Power Plant (coal), 18% Power Plant (natural gas), 9% large Combined Heat and Power plant (natural gas), 2% large Combined Heat and Power plant (coal), 16% small Combined Heat and Power plant (natural gas) and 5% electric boiler.

The marginal electricity production in Life Cycle Assessments is normally either coal or natural gas (Lund et al., 2009), accordingly these have been used for the sensitivity analysis.

A.3.7 Summary of the Life Cycle Inventory Data

Table A.11 presents an overview of the life cycle inventory data used in this project as regarding the storage of slurry, for both pigs and cows’ slurry.

The inputs to the processes are 1000 kg of slurry “ex housing”. All emissions and consumptions are calculated relative to this 1000 kg of slurry going into the process.

Table A.11 Life cycle data for storage of slurry (reference scenario). All data per 1000 kg of slurry “ex housing”.

A.4 Transport to field

A.4.1 System boundaries and description of the process “Transport to field”

The transport of slurry to the fields can be carried out by a tractor with trailer or by truck. For small distances, it is common to use a tractor with trailer and for long distances, a truck is used. According to Pedersen et al. (2007), the trucks transport capacity is up to 35 m³ per trip.

Transport of the slurry from the slurry tank to the fields is estimated to 10 km, as described in section 3.2. Sensitivity analysis has been made for this assumption with 2 km and 32 km. Data for the transport is based on a mix of data from the Ecoinvent process “Transport, tractor and trailer” (10 km).

For the longer distances in the sensitivity analysis, the transport above 10 km is modelled by the Ecoinvent process “Transport, lorry >32t, EURO3”. The Ecoinvent data includes the production of the tractor, trailer and truck (which is a relatively small amount, at it is divided in proportion to all the transport in the entire lifetime of the vehicles).

Emissions from the slurry to air during transport are assumed to be negligible compared to the emissions in the housing units, during storage and during application of slurry, as these emissions are not included in Poulsen et al. (2001), Nielsen et al. (2008a) or Nielsen et al. (2880b).

A.4.2 Summary of the Life Cycle Inventory Data

Table A.12. Life cycle data for transport to field (reference scenario). All data per 1000 kg of slurry “ex storage”.

A.5 Field processes

A.5.1 System boundaries and description of the process “Field Processes”

The process called “Field processes” includes:

• Application of slurry by trail hose application tanker (including diesel for the tractor and production of tractor).
• Emissions to air during application
• Emissions to air during the following period.
• Emissions to water (leaching of N and P)
• Uptake of N. The slurry content of N is assumed to replace mineral N fertiliser (the degree depends on the type of slurry).
• Uptake of phosphorus. It is assumed that the slurry content of P replaces mineral P fertiliser 1:1.
• Uptake of potassium. It is assumed that the slurry content of K replaces mineral K fertiliser 1:1.
• Storage of carbon in the soil. The C-TOOL model complex will be used for estimating C storage in the soil, using the methods described in Gyldenkærne et al. (2007).

The crops on the fields are not included within the system boundaries, as mentioned in chapter 2 under system boundaries.

The life cycle inventory data for application of slurry is shown in table A.17.

The Ecoinvent database contains no data for spreading slurry by trail hose application tanker. As a proxy, data from the Ecoinvent process “Slurry spreading, by vacuum tanker” has been used. The process includes the diesel for slurry application, construction of the tractor, the slurry tanker and a shed, all divided by their estimated life time and slurry amount in this period. The emissions from the combustion of the diesel in the tractor motor are included. The Ecoinvent process includes a diesel consumption corresponding to 0.25 litre diesel per 1000 kg slurry. The diesel consumption in Ecoinvent is at the same level as the 0.3 litre diesel per 1000 kg slurry estimated for slurry spreading by Dalgaard et al. (2002). Adamsen (2004, table 8) estimates a diesel consumption of 0.67 litres of diesel per 1000 kg of slurry for application of 30 tons of slurry per ha. M Kjelddal (2009) estimates that application of slurry by trail hoses consumes approximately 0.4 litres of diesel per 1000 kg of slurry. The calculations are based on the estimate by M Kjelddal (2009), modelled by adjusting the Ecoinvent data to 0.4 litres of diesel per 1000 kg of slurry.

A.5.2 Emissions of CH₄ and CO₂

The CH₄ emission on the field is assumed to be negligible, as the formation of CH₄ requires an anaerobic environment, which is under normal conditions not is the case in the topsoil.

CO₂ emissions are modeled by the dynamic soil organic matter model C-TOOL(Petersen et al., 2002; Gyldenkærne et al., 2008). The development in organic soil N is modeled by assuming a 10:1 ratio in the C to N development.

A.5.3 Emissions of NH₃, N₂O and other N compounds

Significant amounts of NH₃ are lost in the period after application .The NH₃ emissions occurring after application are based on Hansen et al. (2008) (as recommended by T Birkemose, personal communication January 2009). According to Hansen et al. (2008, page 23), the emissions of NH₃ during the very application corresponds to 0.5% of the TAN for trail hose application. In this context, it is assumed that the amount of TAN (NH₃+ NH₄⁺) is the same as the amount of NH₄⁺ ex storage, which is a reasonable approximation at pH 7.8. NH₄⁺-N “ex storage” is calculated as 79% of the total N for fattening pigs and as 58% of the total N for dairy cows (Hansen et al., 2008). Poulsen et al. (2001, page 130) and DJF (2008b) give an estimate of 75% for pig slurry and 60% for cattle slurry, however, this is not used in their calculations (personal communication, H Damgaard Poulsen, January 2008). As the proportions in Hansen et al. (2008) are based on measurements of more than 500 slurry samples, and as the calculation of the ammonia emissions occurring during application based on data from Hansen et al. (2008), the 79% for pig slurry and 58% for cattle slurry from Hansen et al. (2008) is used in this study for the calculation of NH₃ emissions occurring after application..

It should however be emphasized that there is a huge uncertainty connected to the amount of NH₃ emitted after application. In fact, the NH₃-emissions depend on a variety of factors, e.g. application method, soil type, weather (sun/ overcast sky), temperature, wind speed and height of the crop. Accordingly, it is not possible to identify the “true” emission. The values in Hansen et al. (2008) are based on model calculations verified by measurements. The emission factors for trail hose application for the most typical application times from Hansen et al. (2008, page 33) is shown in table A.13. When there is application to a soil without crop, it is assumed that the slurry is ploughed down after a maximum of 6 hours, as this is required by law in Denmark (Hansen et al., 2008).

Table A.13. NH₃ emissions after application, based on Hansen et al. (2008). Emissions are expressed in NH₃-N loss in percent of NH₄⁺-N content in the slurry at the time of application.

^a Hansen et al. (2008) use a relation between NH4+-N and total-N in slurry of 79% for pig slurry and 58% for cattle slurry.

It is assumed that the slurry is partitioned in the crop rotation as specified in section 3.1. Taking pig slurry application to winter wheat as an example of the calculation, 133.5 (see section 3.1) kg N ha^-¹ in slurry is assumed applied in April. According to Hansen et al. (2008) the loss is 14.8 % of TAN. Assuming a TAN content of 58% of the total N content, the loss from soil and leaves becomes 135 kg N ha^-¹ * 0.58 * 0.148 = 11.46 kg N ha^-¹, in addition to the spreading loss itself of 0.5% of TAN, which is 133.5 kg N ha^-¹ * 0.05 = 0.67 kg N ha^-¹, totalling 12.13 kg N ha^-¹.

When performing an area and slurry-N weighed average of all the losses in the crop rotation, a loss of 0.138 g NH₃-N per g TAN in the pig slurry is obtained. For cattle slurry, the equivalent emission coefficient becomes 0.217 g NH₃-N per g TAN in the slurry. The two latter coefficients includes the spreading loss of 0.5%.

The NH₃ emissions for pig slurry thus become 0.50 kg NH₃-N per 1000 kg pig slurry ¹⁹ and 0.75 kg NH₃-N per 1000 kg cattle slurry ²⁰

It should be emphasized that there are significant uncertainties related to the NH₃ emissions. Nielsen et al. (2008b) use an emission factor of 5% of N “ex storage” as basis for the Annual Danish Emission Inventory Report to UNECE . In Sommer and Hansen (2004) it is shown that for cattle slurry the NH₃ emission can vary from 4- 26% of the NH₄⁺-N “ex storage” (trail hose application tanker on winter wheat). For pig slurry the NH₃ emissions was in the range of 3-18% NH₄⁺-N “ex storage” (trail hose application tanker on winter wheat). There were huge variations due to season, temperature and height of the crop. Kai et al. (2008) found a significantly higher emission factor: The NH₃ emission corresponded to almost 50% of the applied NH₄⁺-N “ex storage” during a 7 days period (pig slurry applied by trail hose application tanker on sandy loam soil with 5% clay).

A.5.4 Emissions of N₂O and NO_X

The direct N₂O emissions are 0.01 kg N₂O-N per kg N “ex storage” for application of animal wastes to soil, based on IPPC (2006, table 11.1). The uncertainty range is 0.003 - 0.03 kg N₂O-N per kg N “ex storage”.

In addition, the indirect N₂O emissions have been included in accordance with the IPCC (2006) guidelines, as described in section A.2, i.e. 0.01 kg N₂O–N per kg (NH₃–N + NO_X–N volatilised) (IPCC, 2006, table 11.3).

Nitrate leaching also lead to indirect N₂O emission, corresponding to 0.0075 kg N₂O–N per kg N leaching/runoff (IPCC, 2006, table 11.3).

The emissions of NO and NO₂ are combined as NO_X-emissions, as separate data on NO and NO₂ has not been available. According to Nemecek and Kägi (2007) (page 36) the NO_X emissions can be estimated as: NO_X = 0.21 * N₂O. When taking the molar weights into consideration (assuming NO_X = NO₂) this corresponds to NO_X–N = 0.1 * N₂O-N. It is considered that it is a “rough expert estimate”, however, as the relative contribution has minor significance for the overall results, it is considered to be adequate.

The N₂ emissions are based on the estimates from SimDen (Vinther, 2004).

For soil type JB3 the N₂-N:N₂O-N ratio is 3:1 and for soil type JB6 the N₂-N:N₂O-N ratio is 6.

A.5.5 Nitrogen leaching

Once applied to the field, the N from both mineral and animal fertiliser is assumed to have a limited number of fates: ammonia volatilization, emission from nitrous gasses, removal by harvest, incorporation in the pool of soil organic matter and finally nitrogen leaching. Nitrogen leaching is predominantly in the form of nitrate-N, but may also occur in the form of ammonia and organic N. In a Danish context, erosion losses of N can largely be ignored. The crop net ammonia exchange is very small and uncertain.

Utilizing consequential LCA, the focus is on marginal changes, which in the case of the soil-crop system with focus on slurry means changes in the amount of applied mineral N and slurry composition changes.

To illustrate this, Figure A.2 shows the postulated response of a grain crop to available mineral N.

Figure A.2. N flows at different levels of mineral N fertilisation. This figure is meant to illustrate a general response, and does not address a specific crop.

The responses are non-linear, and it falls out of the scope of this project to estimate all the relevant response curves. But for this purpose, many of the responses may be treated as linear, because they have rather small amplitudes. Take the harvested N in figure A.2 as an example. The value of this will only change slightly due to the different slurry treatments.

So what need to be determined here are mainly linear slopes within small intervals.

Looking again at figure A.2, it is apparent that the leaching loss curve to a large extend is the inverse of the N harvested curve, and under predominate Danish conditions these are the two major fates of field N input. So an essential precondition for good N leaching response estimates is good estimates for N uptake responses in the harvested part of the crop. Unfortunately, the two dynamic models developed for Danish conditions (Daisy, Hansen et al., 1991 and FASSET, Berntsen et al., 2003) are not at present refined to a point where they give very accurate responses of N yield to applied N. One of the problems related to this, is that the N yield at no N application for both models is significantly lower than typical measured values, which crudely put “twists” the entire N response curve for both models. For a more comprehensive discussion on the use of dynamic models, see Petersen et al. (2007).

The FarmN model, derived on the basis of the recommendations in (Petersen et al., 2007) is developed with the aim of giving robust N emission estimates at average yield levels, but does not have an explicit N yield curve, and therefore is not suited for this specific purpose either, because of the pivotal importance of the marginal N responses of the fates of figure A.2.

For the present purpose, we therefore take basis in measured N yield responses, where Landscentret has performed a very large number of field trials throughout Denmark, with different levels of added mineral N.

Pedersen (2008) provides an overview of the national field trials for the later years. This study draws on data from the period 19988 – 2008, with data kindly provided by Leif Knudsen, Landscentret. We utilize information on winter wheat and spring barley, as these are the most common grain crops. At the level of norm fertilisation, the grain yield N recovers approx 36.6 % of the added N on JB3 and 39.9 % on JB6. At lower N levels, the recovery is up to approx. 40% and 50 %, respectively, supporting the non-linear response outlined in figure A.2. For spring barley, the corresponding recoveries are approx. 23.8 % (JB3) and 27.5 % (JB6), with recoveries up to respectively 43 % and 36 %, respectively, at lower N levels. These values are calulated by taking the grain N yield response (3^rd order polynomium approximation, corrected for N carry-over effect of the previous crop) at the norm N fertilisation level.

For simplicity, it is assumed that 50% of the available straw is bailed, and 50% left on the field. This may differ a lot from region to region and farm type to farm type, though. At dairy farms a lot of straw would typically be bailed for own use, while at pig farms the straw would typically either be left on the field or sold. According to Danmarks Statistik (2008), 43 % of the straw in Denmark is left on the field. For both crops, the available straw constitutes 23 % of the dry matter in grain, after Gyldenkærne et al. (2007, Table A1).

According to Landbrugets Rådgivningscenter (2005) the protein concentration in wheat straw is approx. 29% of the concentration in grain. By harvesting the above amount of straw, the marginal harvested N rises to 39.1 % (JB3) and 42.6 % (JB6). The protein concentration in barley straw is approx. 37 % of the grain concentration, whereby the marginal N harvest rises to 25.8 % (JB3) and 29.9 % (JB6). For grain-rich crop rotations, we use the average values of winter wheat and spring barley, giving recoveries of 32.5 % (JB3) and 36.2 % (JB6). Bearing in mind that the responses are obtained from 1-year trials, a significant part of the N in plant residues, and to some extent possible mineral N remainders in the soil after harvest will be available for the following crops. It is crudely assumed that 50 % of the surplus N is lost by leaching after harvest, both caused by mineral N remainders and rapid initial N mineralization in autumn.

The majority of the remaining 50 % is assumed to become available by mineralisation of organic bound N. Therefore the plant uptake “value” of mineralized N versus N from mineral N in fertilizer must be determined. In Petersen et al. (2006, Appendix H), the plant uptake value of mineralized N relative to mineral fertilizer is an average of 65.3 % on JB3 and 73.0 % on JB6. So a significant proportion of the mineralized N is re-utilised by the subsequent crops. Correcting the above marginal harvest N estimates for this effect we obtain 32.5 + 32.5 * 65.3/100 * 50/100 = 43.1 % for a grain-rich crop rotation on JB3, and 36.2 + 36.2 * 73.0/100 * 50/100 = 49.4 % for a grain-rich crop rotation on JB6.

The magnitude of this “carry-over” effect of N fertilizer level is in concordance with winter wheat results from Thomsen et al. (2003), where a long-term previous fertilizer level difference of 78 kg N ha^-¹ on clayey soil resulted in an extra harvest of approx. 10 kg N ha^-¹, where the present coefficients also would predict 10 kg N ha^-¹. Note though that these changes are so minute relative to the uncertainties associated with such field experiments, that this apparent concordance should not be overemphasized.

Calculating the soil N changes with C-TOOL , an additional 9.6 % are incorporated into the soil N pool by adding extra mineral N. This is done by taking the grain yield response (3^rd order polynomium approximation, corrected for carry-over effect of the previous crop) at norm N fertilisation for respectively wheat and barley, and utilize the allometric functions for C crop distributions from Gyldenkærne et al. (2007). Hereby the residue increases by an infinitesimal small increase in N fertilisation with 50 % removal of straw can be calculated. The values are averages for wheat and barley on the respective soil types.

The above leads to the first set of marginal responses in Table A.14.

Table A.14. The 10-year fate of a small change in mineral n application at normal fertilisation levels in a grain-rich crop rotation. Numbers for 100 years in parentheses (see text further below).

When comparing with the quite similar table 21 in Petersen & Djuurhus (2004), the values for marginal leaching is higher in this table, compared to the 0.25 – 0.35 in Petersen & Djuurhus (2004). The lowest value is for sandy soils, and the highest for clay soils. These values are approximately the values obtained by using the NLES₃ N leaching model (Kristensen et al., 2003) though, whereas the present leaching is based on a mass conservation principle.

The other model, besides NLES₃, utilised in the final evaluation (Grant and Waagepetersen, 2003) of the “Vandmiljøplan II” (the Danish Action Plan on the Aquatic Environment II) is SKEP/Daisy (REF). This model gives a higher marginal response for mineral N of 0.59 on JB3 soil, whereas it gives a slightly lower response of 0.27 on JB6 soil (both from Knudsen and Østergard, 2005). Taking a crude average of the response for NLES₃ and SKEP/Daisy for JB3 and JB6 gives an approximate average marginal response of 0.37, which, although unintended, is identical to the average response the present approach yields.

The analogous N fate responses for pig and cattle slurry N may also be obtained.

First the ammonia and denitrification losses are taken from table A.17.

Thereafter the harvested N in bailed straw and grain is calculated by utilizing the substitution values of 0.75 and 0.7, for respectively pig and cattle slurry. As there for cattle slurry, after all gaseous losses, is 71.1 % (JB3) resp. 67.3 % (JB6) of the applied N left, this assumed utilization for plant uptake appears high relative to the sum of organic and mineral N present according to these estimates, before soil N incorporation. It falls out of the scope of this study to estimate possibly improved coefficients for crop availability though, so the present estimates are utilised.

The soil N changes for slurry amendment are calculated with the C-TOOL model.

Finally, the leaching response is calculated by the mass conservation principle.

Table A.15. The 10-year fate of a small change in pig slurry n application at normal fertilisation levels in a grain-rich crop rotation. Numbers for 100 years in parentheses (see text further below).

Table A.16. The 10-year fate of a small change in cattle slurry n application at normal fertilisation levels in a grain-rich crop rotation. Numbers for 100 years in parentheses (see text further below).

One of the overlooked challenges of determining marginal responses is the significance of the considered time scale. When for instance adding extra slurry to the soil, with full substitution for the fertiliser value in terms of applied mineral N, the different fates of N will change radically over the decades, as figure A.3 exemplifies.

Figure A.3 Dependence of the average change in N partitioning on the averaging period, taken from Petersen et al. (2005). The curves represent the average annual difference from the onset of the simulation between continuous slurry application and mineral fertilisation, calculated with the FASSET (Berntsen et al., 2003) agroecosystem model..

Note that the changes in figure A.3. are estimated consequences of a lasting and big change, relative to the baseline scenario, whilst the present estimates are for a “one-event” extra addition of a minute amount of mineral or organic fertiliser. So the time-related changes in respectively figure A.3. and in the present study are not completely comparable.

When transforming the above 10-year considerations to 100-year values, the additional mineralisation of N is calculated first, utilising C-TOOL. The mineralized N is assumed by IPCC to be subject to denitrification, with the same factor as for N amendment.

The N for harvest from mineralization, relative to applied mineral N, is calculated with the same factors as utilized for constructing table A.14.

Subsequently, the new value for leaching may be calculated. This is done for the mineral N application, and the two slurry types.

The responses derived here are presumed well suited for the grain-rich crop rotation of the pig farm. It is also presumed that the responses are valid for the cattle farm crop composition, but the grass-clover mixture occurring here adds another level of complexity to the system and its responses. Unfortunately this is also a system which has been investigated less. So the coefficients are more uncertain for this crop composition.

A.5.6 Phosphorus leaching

The loss of phosphorous from fields is affected by complex dynamics influenced by the soil phosphorus levels, climate, topography, soil conditions, crop type and method of cultivation of the fields. Hence, there are tremendous variations in the loss of phosphorous from different fields (Poulsen and Rubæk, 2005). The routes for the agricultural phosphorous loss are many, such as erosion, surface runoff, leaching to drains, contribution from the surface-near groundwater etc. According to Poulsen and Rubæk (2005), erosion-based losses are just over 50% of the estimated total losses, and bank erosion is by far the largest individual contributor. Leaching from drained wetlands is another significant source.

Even though there is no clear connection between the input of phosphorous to fields and the leaching of phosphorous, a continued net input of surplus phosphorous to agricultural farming soil will - all things being equal – lead to increased risk of loss of phosphorus from the agricultural land (Poulsen and Rubæk, 2005). The loss of particle bound phosphorous by surface runoff and leaching to drains increases with increasing content of phosphorous in the soil. Moreover, the risk of leaching of dissolved phosphorous increases with increasing phosphorous saturation of the soil as the ability of the soil to retain the phosphorous decrease.

It should be emphasized that there is no linearity between the phosphorous added to a field and the leaching of phosphorous. Accordingly, the estimates in this study should only be seen as very rough estimates! Detailed modelling of the phosphorous leaching is beyond the scope and budget of this project.

Poulsen and Rubæk (2005) give some rough estimates for the leaching of agricultural phosphorus in Denmark (data for year 2000):

• The agricultural losses of phosphorous to the aquatic environment ranges between 690 and 1300 tons P per year depending on the method, the time-scale and the input data used according to the national monitoring programme, NOVA (Poulsen and Rubæk (2005) page 28).
• At the national level, phosphorus excretion in animal manure totalled 55000 tons P (year 2000 level) (Poulsen and Rubæk (2005) page 19)
• The input from mineral fertilisers was 17300 tons P per year (Poulsen and Rubæk (2005) page 19).
• Inputs from waste, incl. sewage sludge and atmospheric contribution were in the range of 5800 tons P per year (Poulsen and Rubæk (2005) page 19).

Based on data from Poulsen and Rubæk (2005), the leaching of phosphorous in Denmark corresponds to 1.2% of the P input to the field ²¹. It should be emphasised that this should be regarded as a very rough estimate, and that Poulsen and Rubæk (2005) would probably not use their data for this calculation themselves, as the leaching of phosphorous is caused by years and years application of surplus phosphorus to the soil and not the consequence of the application one year. As the application of P to soil has been significantly higher during the last 50-60 years (Poulsen and Rubæk, 2005, page 35 figure 1.2), an excess amount of phosphorous has been build up in the soil, and the leaching is reflecting the current soil phosphorus levels rather than the input in one year. When taking this into account, the leaching of phosphorous is significantly lower than the 1.2% of the phosphorous input to agricultural soil. Accordingly, the 1.2% should be regarded as an estimate for the maximum.

Since 2000, the contribution of phosphorus from manure has been reduced significantly (Vinther and Poulsen, 2008). The yearly input from mineral fertilisers, organic waste and from animal feed (leading to P in manure) has decreased by 4000 tons P from 2001/2002 to 2007/2008 and the yearly output by plant products and animal products has increased by 1000 tons and 1500 tons, leading to a total decrease in the surplus P of 6500 tons. Even though the surplus phosphorus has decreased since 2000, the data from Poulsen and Rubæk (2005) has been used for estimating the relationship between the phosphorous applied to the field and the agricultural leaching of P, as newer data has not been available.

It could be discussed whether the leaching of phosphorus should be seen as a percentage of the total input of P to the field, or in relation to the surplus amount of P applied to field (i.e. the input of P minus the uptake of P by the plants). Ideally the P balances in this study should be based modelling of field balances including P added as manure, P added as mineral fertiliser and P removed with the crop harvested, as done for nitrogen leaching in section A.5.5, and in addition to this inclusion of modelling of soil phosphorus levels. However, it has not been possible to model the phosphorous leaching, as it is far more complicated than modelling of nitrogen (which is not simple either). As mentioned above, Poulsen and Rubæk (2005) describe the complexity of modelling P leaching.

Nielsen and Wenzel (2005) assume a leaching of phosphorous is in the order of 5% of the net surplus application, assuming that phosphorus spread with manure on farmland at farms with 1.4 lifestock units²² per ha or more is on average in the order of 30 kg P per ha while the plant uptake is about 20 kg P per ha (Kronvang et al., 2001). The net surplus is then about 10 kg P per ha. Nielsen and Wenzel (2005) emphasize that it is a rough estimate and perform sensitivity analysis for a leaching of 0% and 100% of the net surplus. Dalgaard et al. (2006) assume that 2.9% of the farm gate P balance leached as phosphate.

If the modelling is based on the net surplus amount of phosphorous, it implies that if there is no surplus of P applied to the field, there will be no leaching of phosphorous. This is not necessarily true. The erosion based loss mihgt occur and as long as the soil contains phosphorous, there will probably be a small loss, even if P is not applied to the field in excess amounts. When watching the leaching of N as a function of the amount of applied N in figure A.2 in the beginning of section A.5.5 for nitrogen leaching, it can be seen that the there is a small leaching of N regardless of the amount of N applied (for small amounts of applied N). For higher amounts of applied N, the N leaching increases with increasing amounts of applied N. The leaching of P might follow a similar pattern.

In this study, the EDIP 2003 approach for phosphorous leaching has been applied (Hauschild and Potting, 2005). It builds on a simple linear assumption, which will most likely not be applicable for all levels of application of P. However, as the application of phosphorous to field is the same for the reference system and the new technologies in this study, it has no consequences for the comparisons (however, this only applies for the new technologies in this study). For future assessments of new slurry management technologies, where the amount of P applied to field is changed, sensitivity analysis should be carried out, applying different approaches for phosphorous leaching modelling – based on the total input of P to field (as done in this study) and based on the surplus amount of P applied to field.

According to Hauschild and Potting (2005, Annex 6.3), 10% of the P applied to field has the possibility of leaching (this is the amount of P that should be entered into the life cycle modelling in SimaPro). Of this, 6% actually reach the aquatic recipients according to the model used by Hauschild and Potting (2005). As a result, 0.6% of the amounts of P applied to field actually reach aquatic recipients according to the EDIP 2003 method. When keeping the huge uncertainty on the estimates in mind, the estimate by Hauschild and Potting (2005) (i.e. that 0.6% of the P input to field reach aquatic recipients) is at the same level as the estimates for the phosphorous leaching to the aquatic environment based on Poulsen and Rubæk (2005) (1.2% as mentioned above).

It is assumed that the leaching of P from mineral P fertilisers is the same as leaching of P from slurry.

The amount of P applied to field is 1.04 kg per 1000 kg pig slurry ex storage and 0.98 kg per 1000 kg dairy cow slurry ex storage, see table A.1 and A.2. 10% of this has the potential of leaching.

A.5.7 Summary of the Life Cycle Inventory Data

Table A.17. Life cycle data for application of slurry and field processes (reference scenario). All data expressed per 1000 kg of slurry ex outdoor storage.

A.6 Avoided mineral fertilisers

The application of N, P and K in the slurry replaces mineral fertilisers.

The replaced production and application of mineral fertilisers are subtracted from the reference scenario. When taking the consequential, marginal approach into consideration, the production of mineral N fertilisers that are affected by the slurry application should be identified, i.e. the production of mineral fertiliser that will affected when the N in pig slurry and cattle slurry is used more efficiently. Furthermore, the affected production of P and K fertiliser should be identified.

A.6.1 Amount of replaced mineral fertilisers

In Denmark, the farmers’ use of N fertilisers is restricted by Danish law (Gødskningsbekendtgørelsen, 2008, and Gødskningsloven, 2006). It means that the amount of N fertiliser farmers are allowed to bring out has an upper ceiling, both as mineral fertiliser and animal slurry. The farmers have to make accounts on their fertiliser use, and they have to include a fixed amount of the N content of the animal slurry in their fertiliser accounts. The substitution requirements in the Danish law are 75% for pig slurry and 70% for cattle slurry. The requirement means, that when the farmer brings out 100 kg total-N in pig slurry, he has to include it in the fertiliser account corresponding to 75 kg N in mineral fertiliser which means that the farmer has to reduce the consumption of mineral N fertiliser by 75 kg. Accordingly, it is assumed that 100 kg N added in the pig slurry replaces 75 kg N in mineral fertiliser. 100 kg N added in the cattle slurry replace 70 kg N in mineral fertiliser.

As the farmer calculates the amount of N according to the Norm Data, the “avoided mineral N fertiliser” is calculated in accordance with the N ex storage from the Norm data, i.e. without the loss of N due to N₂O, NO_X and N₂ emissions. These are shown in table A.1 for pig slurry and table A.2 for dairy cow slurry. Accordingly, for pig slurry the avoided mineral N fertiliser is calculated as: 75% of 5.00 kg N ex storage = 3.75 kg N (i.e. not calculated as 75% of the calculated N content in this study as 4.80 kg N, see table A.1). For dairy cow slurry the avoided N in mineral fertiliser corresponds to 70% of 6.02 kg N (Norm Data ex storage) = 4.21 kg N. The avoided N in mineral fertiliser is higher when calculated according to the Norm Data ex storage than if calculated as the percentage of the data in this study. However, when modelling the consequences of what the farmer does, it is the Norm Data the farmer uses for his N accounts.

For P and K the conditions are different.

The fertiliser value of P and K applied in slurry is generally assumed to have the same value as in mineral fertilisers (i.e. that 1 kg P in slurry has the same plant availability as 1 kg P in mineral fertilisers). This assumption is also used by e.g. Thyø and Wenzel (2007) and Börjesson and Berglund (2007). Poulsen and Rubæk (2005, page 26) support this assumption for phosphorous “In terms of plant nutrition, phosphorus in readily soluble mineral fertilizer and animal manure is considered of equal value. It appears though that animal manure phosphorus is more mobile and is more easily transported to deeper soil layers than mineral fertilizer phosphorus”.  Sommer et al. (2008) mention that the immediate availability of P in slurry may be somewhat lower than for mineral fertilisers.

Nevertheless, the assumption that the P and K fertiliser value is the same for slurry and mineral fertilisers does not necessarily mean that that the replacement is carried out 1:1. In consequential life cycle assessments, the key issue is to identify the consequences in real life. In this case, the question is: “What would the consequence be if the farmer did not apply slurry and thereby P and K to the field?”. The answer is that he would apply mineral fertilisers instead. However, the application of slurry to field leads to excess amounts of P and K when the slurry is applied in accordance with the requirements set by Danish law (Miljøministeriet, 2006). If applying mineral fertilisers instead of slurry, he would probably not apply P and K in excess amounts. Poulsen and Rubæk (2005) assumes that mineral P fertiliser is applied in amounts that are adjusted to the soil phosphorus levels and the needs of the crop ²³. Accordingly, it is assumed that P and K would not be applied in excess amounts if mineral fertilisers where used, and as a result, P and K in slurry does not replace mineral fertiliser 1:1.

In this study, it is assumed that the amount of P and K applied as mineral fertilisers is based on measurements of the needs of the crop. It is assumed that this does not lead to application of P and K in excess amounts. Accordingly, only part of the P and K applied in the slurry actually replace mineral fertilisers.

The calculations of the replacement of P and K mineral fertilisers is based on the requirements set by Danish Law (Miljøministeriet, 2006) and the recommendations for fertilising crops by Plantedirektoratet (2008).

The amount of slurry applied to the field is calculated in accordance with Danish Law (Miljøministeriet, 2006), i.e. 1.4 livestock units²⁴ per ha for pigs and 1.7 livestock units per ha for cattle. There is 0.85 dairy cow per livestock unit (heavy race) and 35 fattening pigs per livestock unit (Miljøministeriet, 2006). Note that the amount of animals per livestock units were defined in accordance with the Norm Data from 2000 for fattening pigs and 1999 for dairy cows and as the Norm data (N ex storage per animal) has changed, 1 livestock unit does not correspond to 100 kg N ex storage anymore, even though this was the original definition. The amount of slurry and content of N, P and K is given by DJF (2008), which is in accordance with the guidelines from Plantedirektoratet (2008).

For fattening pigs, the amount of slurry applied to 1 ha is: 1.4 livestock units per ha * 35 fattening pigs per livestock unit * 0.52 tonnes slurry per pig (ex storage) = 25.48 tonnes pig slurry per ha. This amount of slurry contains:

• 5.00 kg N per tonnes slurry * 25.48 tonnes slurry = 127.4 kg N per ha
• 1.04 kg P per tonnes slurry * 25.48 tonnes slurry = 26.50 kg P per ha
• 2.60 kg K per tonnes slurry * 25.48 tonnes slurry = 66.25 kg K per ha

For dairy cows, the amount of slurry applied to 1 ha is: 1.7 livestock units per ha * 0.85 dairy cows per livestock unit * 21.3 tonnes slurry per dairy cow (ex storage) = 30.78 tonnes dairy cow slurry per ha. This amount of slurry contains:

• 6.02 kg N per tonnes slurry * 30.78 tonnes slurry = 185.3 kg N per ha
• 0.98 kg P per tonnes slurry * 30.78 tonnes slurry = 30.16 kg P per ha
• 5.65 kg K per tonnes slurry * 30.78 tonnes slurry = 173.91 kg K per ha

The amounts of P and K recommended to each crop type is based on the recommendations given by Plantedirektoratet (2008, table 1), calculated in accordance with the crop rotation given in section 3.1.

The 6 years crop rotation for pig slurry is defined in section 3.1 to: winter barley – winter rape – winter wheat – winter wheat – spring barley with catch crop – spring barley. With this crop rotation, the recommended amounts of P and K are 21.5 kg P per ha and 64 kg K per ha (average, weighted with regard of the crop rotation).

The 5 years crop rotation for dairy cow slurry is defined in section 3.1 to: whole crop silage – grass clover mixture – grass clover mixture – spring barley with catch crop – spring barley. With this crop rotation, the recommended amounts of P and K are 27.8 kg P per ha and 125.8 kg K per ha (average, weighted with regard of the crop rotation).

In addition to the P and K in the slurry, mineral fertilisers might be added by the farmer. However, this is not caused by the slurry management and is not relevant for the goal of this study. If the farmer adds more mineral P and K than the recommended amounts, they are either “excess amounts” or due to that the P or K from the slurry is not available for the crop. It is beyond the scope of this study to model the faith and availability of P and K at the field.

The excess amount of P and K and the percentage of the P and K in the slurry that actually replace mineral fertilisers are calculated:

For fattening pig slurry:

• 26.50 kg P per ha is added in the slurry, 21.5 kg P per ha is recommended, excess amount is 5 kg P per ha, corresponding to that only 81.1% of the P in the slurry replace mineral fertilisers in the reference scenario.
• 66.25 kg K per ha is added in the slurry, 64 kg K per ha is recommended, excess amount is 2.25 kg K per ha, corresponding to that only 96.6% of the K in the slurry replace mineral fertilisers in the reference scenario.

For dairy cow slurry:

• 30.16 kg P per ha is added in the slurry, 27.8 kg P per ha is recommended, excess amount is 2.36 kg P per ha, corresponding to that 92.2% of the P in the slurry replace mineral fertilisers in the reference scenario.
• 173.91 kg K per ha is added in the slurry, 125.8 kg K per ha is recommended, excess amount is 48.11 kg K per ha, corresponding to that only 72.3% of the K in the slurry replace mineral fertilisers in the reference scenario.

Accordingly, the replaced amounts of mineral fertilisers for pig slurry in the reference system are calculated relative to the “functional unit” i.e. 1000 kg slurry ex animal:

• Mineral N fertiliser: 5.00 kg N per 1000 kg slurry ex storage [the value given by the Danish Norm Data, as explained in table A.1] * 1086 kg slurry ex storage per 1000 kg slurry ex animal * 75% [the replacement value for pig slurry according to (Gødskningsbekendtgørelsen, 2008] = 4.073 kg mineral N fertiliser
• Mineral P fertiliser: 1.04 kg P per 1000 kg slurry ex storage * 1086 kg slurry ex storage per 1000 kg slurry ex animal * 81.1% = 0.916 kg P
• Mineral K fertiliser: 2.60 kg K per 1000 kg slurry ex storage * 1086 kg slurry ex storage per 1000 kg slurry ex animal * 96.6% = 2.73 kg K

The replaced amounts of mineral fertilisers for dairy slurry in the reference system are:

• Mineral N fertiliser: 6.02 kg N per 1000 kg slurry ex storage [the value given by the Danish Norm Data, as explained in table A.2] * 1044 kg slurry ex storage per 1000 kg slurry ex animal * 70% [the replacement value for cattle slurry according to (Gødskningsbekendtgørelsen, 2008] = 4.399 kg mineral N fertiliser
• Mineral P fertiliser: 0.98kg P per 1000 kg slurry ex storage * 1044 kg slurry ex storage per 1000 kg slurry ex animal * 92.2% = 0.943 kg P
• Mineral K fertiliser: 5.65kg K per 1000 kg slurry ex storage * 1044 kg slurry ex storage per 1000 kg slurry ex animal *72.3% = 4.26 kg K

A.6.2 Market considerations for mineral fertilisers

Mineral fertilisers are traded on markets. According to the consequential methodology, the fertilizers (and the technology used to produce them) affected by a change of the slurry management in Denmark will only be the marginal ones. According to Weidema (2003), the technology with the lowest long-term production cost is the one that is likely to be implemented in case of increasing demand and would therefore be the marginal. Oppositely, if the demand trend is decreasing, the least competitive technology is the most likely to be phased out and would represent the marginal in that case.

According to this methodology, it should therefore be known if the demand for mineral fertilisers is an increasing or decreasing trend.This depends on the market considered, i.e. if Europe is regarded as a closed market or whether if it is regarded as an open market.It is however beyond the frameworks of this study to conduct a market analysis for mineral fertilisers and to collect data for old and modern technologies for the production of various types of mineral fertilisers. Accordingly, a pragmatic approach has been taken.

First, the fertiliser’s data available in the Ecoinvent database were taken into account. A variety of fertilizers data is available in this database, as shown below :

It can be noticed from this overview of the available data in EcoInvent as regarding fertilizers that data are presented for both mixed and non mixed fertililizers. For simplification and due to lack of data in the statistics, mixed fertilisers (i.e. fertilisers including a combination of N and P or N, P and K) have not been used as the marginal fertiliser in this study. In fact, the data by Plantedirektoratet (2008) are not detailed enough for estimating the combination of mixed fertilisers as the P and K content are not stated for the mixed mineral fertilisers.

As a second step of the pragmatic approach adopted, the market context for mineral fertilisers was briefly examined, for both N, P and K fertilisers.

The total sale of mineral fertilisers in Denmark is slightly decreasing. However, the worldwide demand for fertilisers is increasing. Alley and Spargo (2007) states: “Growth in the economies of China and India, in particular, as well as other countries has created a greater worldwide demand for fertilisers, and increased use of corn for ethanol production is increasing fertiliser demand in the United States due to expected increases in corn acres. World production capacity for N and P fertilisers is slightly greater than demand while potash capacity is significantly greater than demand, but production has been constrained by several factors in recent years. Adequate supplies of fertilisers appear to be available in the world market, but logistical challenges exist for nitrogen and potash in particular.”

According to table 2 in Alley and Spargo (2007) the worldwide consumption of fertilisers have been increasing significantly from 2004-2007 for N-fertilisers, and P-fertilisers as well as for K-fertilisers.

A.6.3 Marginal N fertiliser

For N fertilisers, the main consumption in Denmark is “N fertilisers mixed with sulphur” and NPK, NP or NK fertilisers according to the statistics by Plantedirektoratet (2008). However, as mentioned above, since the composition of the mixed N fertilisers is not stated in the statistics, it is not possible to use these for the modelling. According to the statistics by IFA (2008), Plantedirektoratet (2008) and Nielsen et al. (2008), the most commonly used non-mixed N fertiliser is calcium ammonium nitrate. The Danish consumption of calcium ammonium nitrate has been decreasing in the period of 1998-2006 (IFA, 2008 and Plantedirektoratet, 2008), which is, however, the case for most of the N fertilisers but ammonium sulphate, “liquid fertilisers” and urea. Alley and Spargo (2007) states that urea is the most widely traded N source in the world. However, it is not very used in Denmark. When analysing the production of calcium ammonium nitrate in the Ecoinvent Database, it is very similar to the production of ammonium nitrate plus limestone (for the calcium). Accordingly, ammonium nitrate is assumed to be the marginal N fertiliser in this study. For the sensitivity analysis, ammonium sulphate has been used as the environmental profile for this is rather different from the impacts of calcium ammonium nitrate in order to assess the significance of this for the overall results.

Accordingly, in this study the process “Ammonium nitrate, as N, at regional storehouse/RER U” from Ecoinvent database has been used. However, the Ecoinvent data has been modified slightly. According to the background documentation for this process (Nemecek and Kägi, 2007), the emission of N₂O from the processing of nitric acid is based on literature data from 1997, which means that the data are probably more than 10 years old. As the N₂O emission from this process has great significant for the overall results of the Life Cycle Assessment in this report, newer data for the N₂O emissions has been applied. According to the BREF document for production of nitric acid, the emission level for N₂O for new plants is 0.12 – 0.6 kg/tonne 100 % HNO₃ and 0.12 – 1.85 kg/tonne 100 % HNO₃ for existing plants (European Commission, 2007). However, the BREF document states: “Industry and one Member State claim that the BAT range should include 2.5 kg N₂O/tonne 100 % HNO₃ for existing plants.” (European Commission (2007), section 3.5, page 140). In the Ecoinvent database, the N₂O emission for the production of nitric acid corresponds to 8.39 kg N₂O/tonne 100 % HNO₃. The N₂O emission has been modified to 2.5 kg N₂O/tonne 100 % HNO₃ in this study.

The application of mineral fertilisers are included by the Ecoinvent process ”Fertilising, by broadcaster”. Emissions from the diesel consumption by the tractor are included in the Ecoinvent process “Fertilising, by broadcaster”. The diesel consumption in the Ecoinvent data corresponds to a consumption of 6.3 litres of diesel per ha. According to Dalgaard et al. (2002) the energy consumption for application of mineral fertiliser is 2 litres of diesel per ha. In the calculations, the energy consumption by Dalgaard et al. (2002) has been used, modelled by the relative ratio using the Ecoinvent data. In the calculations it is assumed that there is applied 30 tons slurry per ha (i.e. 1/30 of ha per 1000 kg slurry). Accordingly, there is avoided application of mineral fertilisers on an area of 1/30 ha per 1000 kg slurry. A rough estimate is 0.007 litres of diesel per kg fertiliser (assuming that all three fertilisers are applied at the same time).

The application of N fertiliser will lead to emissions of NH₃, which means that by avoiding application of N fertiliser also avoid NH₃ emissions.

Nielsen et al. (2008b) gives a table of the NH₃ emission factors for mineral fertilisers (For calcium ammonium nitrate, the NH₃-N emission factor is 2% of the added N (Nielsen et al. (2008), Annex C, table 2C.6) and European Environment Agency (2007, page B1010-12 table 4.1).

The N₂O emission factor for mineral N fertilisers is 0.01 kg N₂O -N/kg N for application of mineral fertilisers to soil, based on IPCC (2006, table 11.1).

In addition, the indirect N₂O emission have been included in accordance with the IPCC (2006) guidelines, i.e. 0.01 kg N₂O–N per kg (NH₃–N + NO_X–N volatilised) (IPCC, 2006, table 11.3).

For NO emissions from N fertilisers, the emission factor recommended by the European Environment Agency (2007) has been applied: 0.7 % NO-N emissions related to the input of mineral fertiliser N.

A.6.4 Marginal P fertiliser

For P fertilisers, the Ecoinvent database only contains two phosphorous fertilisers that are not “mixed” with either N or K (As described above, it is preferable to model the avoided fertiliser production with non-mixed fertilisers (i.e. not NPK fertilisers, or NP fertilisers), if possible, in order to avoid making a “branch” of avoided processes influencing on each other):

• Single superphosphate, as P2O5, at regional storehouse/RER U
• Triple superphosphate, as P2O5, at regional storehouse/RER U

It is assumed that the avoided P fertiliser is Triple superphosphate. According to Broadley et al. (2006) single superphosphate has generally been replaced by triple superphosphate (at least in the UK).

According to the statistics by Plantedirektoratet (2008), triple superphosphate is more commonly used than single superphosphate. According to the statistics by Plantedirektoratet (2008) it seems that the consumption of triple superphosphate as well as the consumption of single superphosphate has been decreasing since 2001 (however, the two types have been aggregated into one category since 2004 which makes it difficult to distinguish the marked trend for the two types). A sensitivity analysis for this assumption has been carried out as the environmental impacts of producing triple superphosphate is only 65-80% of the environmental impacts of producing single superphosphate according to a screening performed in SimaPro. In this study, the Ecoinvent process “Triple superphosphate, as P2O5, at regional storehouse/RER U” has been used. 1 kg P corresponds to 2.291 kg P₂O₅ ²⁵

According to the background documentation for the Ecoinvent database (Althaus et al., 2007), triple superphosphate is produced from phosphoric acid and phosphate rock. The production of phosphoric acid gives large amounts of phosphogypsum, which is an environmental challenge. The largest productions of phosphoric acid are in U.S. (29%) and Morocco (17%). There are significant differences between the productions and discharges from the U.S phosphoric acid plants and the Morocco plants. In the U.S., the phosphogypsum is discharged from a filter, pumped to decantation basins where the gypsum settles before being recycled. No direct water discharge is assumed due to the closed water circuit. A leaching of 1% is assumed due to rain falling onto the stack area causing leaching into the aquifier. Phosphoric acid plants in Morocco dispose the phosphogypsum directly to the sea where it dissolves, and accordingly, all the phosphogypsum is calculated as short-term emissions to seawater (Althaus et al. (2007) and Anwar and Wissa (unknown year)).

It has not been possible to identify which type of phosphoric acid plant that is the marginal plant. However, in Europe, the emission of phosphogypsum to seawater is no longer accepted in Europe (European Commision, 2007, page 247). In this report, the U.S. phosphoric acid plant is used for the “basic calculation” as this reflects the European production better than the production in Morocco. The Morocco phosphoric acid plant is used for the sensitivity analysis. The environmental profile of the two productions are very similar but for the eutrophication (P) caused by the leaching of phoshogypsum in Morocco, which cause a contribution to “Eutrophication (P)” that is approximately 8.6 times higher for the production in Morocco.

A.6.5 Marginal K fertiliser

The potassium fertilser in this study is modeled by the use of the Ecoinvent process ”Potassium chloride, as K2O, at regional storehouse/RER U”. According to the statistics by Plantedirektoratet (2008), potassium is most commonly applied in mixed mineral fertilisers combined with N, however, it has not been possible to include this due to a lack of data as described above. According to the statistics by IFA(2008) and Plantedirektoratet (2008) potassium chloride is commonly used in Denmark. Plantedirektoratet (2008) assess the yearly consumption of potassium chloride, and there are no data for potassium sulphate. The consumption of potassium chloride in Denmark is slightly fluctuating without a clear increase or decrease. 1 kg K corresponds to 1.205 kg K₂O ²⁶.

A.6.6 Summary of the Life Cycle Inventory Data

The Life Cycle Inventory Data for application of mineral fertilisers are shown in table A.18. The values in the table are positive; however, when the processes are subtracted from the system, they will lead to a “negative contribution” as the emissions are avoided.

Table A.18. Life cycle data for application of mineral fertiliser (reference scenario). All data per 1 kg of mineral fertiliser applied.

[1] The amount of straw corresponds to 2.5 kg (10 kg straw per m² slurry surface (Rasmussen et al. (2001). With a 4 m deep slurry tank this corresponds to 2.5 kg straw per 1000 kg slurry. 2.5 kg straw per 1000 kg slurry = 0.25%

[2] DM in straw: 85% (Poulsen et al., 2001, page 89). Total added amount of dry matter: 2.5 kg * 0.85 = 2.1 kg. This corresponds to 3.4% of the total amount (61 kg in 1000 kg slurry).

[3] As mentioned above, 2.5 kg straw is added to pig slurry during storage.
According to Poulsen et al., 2001, page 89, straw contains 85% dry matter (DM). Straw contains 0.005 kg N per kg DM in straw: N added in straw = 0.005 kg N per kg DM * 0.85 * 2.5 kg = 0.01 kg N (0.2% of the content of N in slurry “ex storage”).
Straw contains 0.00068 kg P per kg DM P added in straw = 0.00068 kg P per kg DM * 0.85 * 2.5 kg = 0.0015 kg P (0.1% of the content of P in slurry “ex storage”).
Straw contains 0.01475 K kg per kg DM per 1000 kg slurry. K in straw = 0.01475 kg K per kg DM * 0.85 * 2.5 kg = 0.031 kg K (1% of the content of K in slurry “ex storage”).

[4] The calculations for slurry from fattening pigs (fully slatted floors) is based on the following assumptions: DM “ex storage” is 61 kg (DJF, 2008a). The DM loss during storage is 5% of DM “ex housing” (DJF (2008b) and Poulsen et al. (2001), i.e. DM(storage loss) = 61 kg * 0.05/(1-0.05) = 3.2 kg DM “ex housing” = DM “ex storage” + DM(storage loss) = 61 kg+3.2 kg = 64.2 kg The DM loss during housing: 10% of DM “ex housing” (DJF (2008b) and Poulsen et al. (2001), i.e. DM(housing loss) = 64.2 kg * 0.10/(1-0.10) = 7.1 kg In Sommer et al. (2001) it is assumed that the loss of DM is identical to the loss of easily degradable VS.
This is also identical to the loss of VS, as VS = VS(easily degradable) + VS(heavy degradable) and as the VS(heavy degradable) is not changed). I.e.: VS “ex animal” = VS “ex storage” + DM(storage loss) + DM(housing loss) VS “ex animal” = 48.8 kg + 3.2 kg + 7.1 kg = 59.1 kg Sommer et al. (2001) assume that 65% of the VS is easily degradable (”ex animal”) for pigs: VS(easily degradable, ex animal) = 59.1 kg * 0.65 = 38.4 kg As the loss of VS(easily degradable) is identical to the DM loss, the calculations are: VS (easily degradable, ex housing) = 38.4 kg – 7.1 kg = 31.3 kg VS (easily degradable, ex storage) = 31.3 kg – 3.2 kg = 28.1 kg Sommer et al. (2001) assumes that 35% of the VS is heavily degradable (”ex animal”) for pigs: VS(easily degradable, ex animal) = 59.1 kg * 0.35 = 20.7 kg. As it is assumed that the heavily degradable VS is unchanged in the housing units and during storage VS (heavily degradable, ex housing) = 20.7 kg and VS (heavily degradable, ex storage) = 20.7 kg.
The calculations for dairy cows follow the same assumptions as for pig slurry.

[5] According to Knudsen and Birkmose (2005), pig slurry contains 38 kg dry matter and 18.2 kg C, i.e. 47.9% of the dry matter is carbon. Cattle slurry contains 64 kg dry matter per 1000 kg slurry and 28.1 kg C, i.e. 43.9% of the dry matter is carbon

[6] According to IPCC (2006), the methane emission can be calculated as:
CH₄ [kg] = VS [kg] * B₀ * 0.67 [kg CH₄ per m³ CH₄] * MCF
The VS amount is “ex animal”.
B₀ = 0.45 m³ CH₄ per kg VS for market swine (IPCC, 2006, Table 10A-7)
B₀ = 0.24 m³ CH₄ per kg VS for dairy cows (IPCC, 2006, Table10A-4)
0.67 is the conversion factor from m³ CH₄ to kilograms CH₄
MCF = 17% for pit storage below animal confinements > 1 month (IPCC, 2006, Table10-17)
MCF = 3% for pit storage below animal confinements < 1 month (IPCC, 2006, Table10-17)
Calculation for fattening pig slurry, containing 64.2 kg VS per 1000 kg slurry (see table A.1 in this report), and assuming that the slurry is in the slurry pits > 1 month:
CH₄ [kg] = 64.2 kg VS per 1000 kg slurry * 0.45 m³ CH₄ per kg VS * 0.67 [kg CH₄ per m³ CH₄] * 0.17 = 3.29 kg CH₄ per 1000 kg pig slurry “ex animal”.
Calculation for dairy cow slurry, containing 104.2 kg VS per 1000 kg slurry (see table A.2 in this report), and assuming that the slurry is in the slurry pits > 1 month:
CH₄ [kg] =104.2 kg VS per 1000 kg slurry * 0.24 m³ CH₄ per kg VS * 0.67 [kg CH₄ per m³ CH₄] * 0.17 = 2.85 kg CH₄ per 1000 kg dairy cow slurry “ex animal”.

[7] For fattening pigs: The carbon loss in the housing units is 3.4 kg (table A.7).
The CH₄ emission is 3.29 kg (see table A.9),, which corresponds to 3.29 kg * 12.011/(12.011 + 4*1.008) = 2.46 kg carbon.
Carbon loss as CO₂-C = 3.4 kg – 2.46 kg = 0.94 kg CO₂-C
0.94 kg CO₂-C which corresponds to 0.94 kg * (12.011 + 2 * 15.9994) / 12.011 = 3.44 kg CO₂

[8] For dairy cows: The carbon loss in the housing units is 5.3 kg (table A.8).
The CH₄ emission is 2.85 kg (see table A.9), which corresponds to 2.85 kg * 12.011/(12.011 + 4*1.008) = 2.13 kg carbon.
Carbon loss as CO₂–C = 5.3 kg – 2.13 kg = 3.17 kg CO₂-C
3.17 kg CO₂-C corresponds to 3.17 kg * (12.011 + 2 * 15.9994) / 12.011 = 11.6 kg CO₂

[9] Kai et al. (2008) that found that an average NH₃–N emission from fattening pig housing units was 0.43 kg per pig produced (95% confidence interval 0.38–0.49 kg NH₃-N). As each pig produces 470 kg slurry the corresponding value is 0.91 g NH₃-N per 1000 kg pig slurry “ex animal”.

[10] Sommer et al. (2006) suggest an emission coefficient of 0.25 kg NH₃-N pr. kg TAN (NH₄⁺-N + NH₃-N) for fattening pigs in housing units with fully slatted floors. Assuming that pig slurry “ex animal” contains 79% NH₄⁺-N (Hansen et al., 2008), and assuming that the pH is 7.8 which means that TAN are almost identical to the NH₄⁺-N amount, the NH₃ emission can be calculated as: 0.25 kg NH₃-N pr. kg TAN * 6.60 kg N per 1000 kg pig slurry “ex animal” (see table A.1) * 0.79 = 1.3 kg NH₃.

[11] Sommer et al. (2006) suggests an emission coefficient of 0.17 kg NH₃-N pr. kg TAN (NH₄⁺-N + NH₃-N) for cattle in cubicle housing units with slatted floors. Assuming that pig slurry “ex animal” contains 58% NH₄⁺-N (Hansen et al., 2008), and assuming that the pH is 7.8 which means that TAN are almost identical to the NH₄⁺-N amount, the NH₃ emission can be calculated as: 0.17 kg NH₃-N pr. kg TAN * 6.87 kg N per 1000 kg pig slurry “ex animal” (see table A.2) * 0.58 = 0.68 kg NH₃.

[12] From IPCC (2006, section 11.2.2): “In addition to the direct emissions of N₂O from managed soils that occur through a direct pathway (i.e., directly from the soils to which N is applied), emissions of N₂O also take place through two indirect pathways.
The first of these pathways is the volatilisation of N as NH₃ and oxides of N (NO_X), and the deposition of these gases and their products NH4+ and NO3
- onto soils and the surface of lakes and other waters. “ and ”The second pathway is the leaching and runoff from land of N from synthetic and organic fertiliser additions, crop residues, mineralisation of N associated with loss of soil C in mineral and drained/managed organic soils through land-use change or management practices, and urine and dung deposition from grazing animals. “ and “The nitrification and denitrification processes transform some of the NH₄⁺ and NO₃ - to N₂O.”

[13] According to IPCC (2006), the methane emission can be calculated as:
CH₄ [kg] = VS [kg] * B₀ * 0.67 [kg CH₄ per m³ CH₄] * MCF
The VS amount is “ex animal”.
B₀ = 0.45 m³ CH₄ per kg VS for market swine (IPCC, 2006, Table 10A-7)
B₀ = 0.24 m³ CH₄ per kg VS for dairy cows (IPCC, 2006, Table10A-4)
0.67 is the conversion factor of m³ CH₄ to kilograms CH₄
MCF = 10% for liquid slurry with natural crust cover, cool climate (IPCC, 2006, Table10-17)
Calculation for fattening pig slurry, containing 64.2 kg VS per 1000 kg slurry (see table A.1 in this report), and assuming that the slurry is in the slurry pits > 1 month:
CH₄ [kg] = 64.2 kg VS per 1000 kg slurry “ex animal” * 0.45 m³ CH₄ per kg VS * 0.67 [kg CH₄ per m³ CH₄] * 0.10 = 1.94 kg CH₄ per 1000 kg pig slurry “ex animal” (which is identical to 1.94 kg CH₄ per 1000 kg pig slurry “ex housing”).
Calculation for dairy cow slurry, containing 104.2 kg VS per 1000 kg slurry (see table A.2 in this report), and assuming that the slurry is in the slurry pits > 1 month:
CH₄ [kg] =104.2 kg VS per 1000 kg slurry “ex animal” * 0.24 m³ CH₄ per kg VS * 0.67 [kg CH₄ per m³ CH₄] * 0.10 = 1.68 kg CH₄ per 1000 kg dairy cow slurry “ex animal” (which is identical to 1.68 kg CH₄ per 1000 kg pig slurry “ex housing”).

[14] Sommer et al. (2001), Appendix 3: 0.7744 kg CO₂-eqv per kg VS excreted / 21 CO₂-eqv per kg CH₄ * 56 kg VS per 1000 kg pig slurry (Sommer et al. (2001, page 44) = 2.07 kg CH₄ per 1000 kg pig slurry.

[15] Sommer et al. (2001), Appendix 3: 0.4222 kg CO₂-eqv per kg VS excreted / 21 CO₂-eqv per kg CH₄ * 80 kg VS per 1000 kg cattle slurry (Sommer et al. (2001, page 44) = 1.61 kg CH₄ per 1000 kg cattle slurry.

[16] For fattening pigs: The carbon loss during storage 1.5 kg (table A.7).
The CH₄ emission is 1.94 kg (see table A.9), which corresponds to 1.94 kg * 12.011/(12.011 + 4*1.008) = 1.45 kg carbon.
Carbon loss as CO₂ = 1.5 kg – 1.45 kg = 0.05 kg C, which corresponds to 0.05 kg * (12.011 + 2 * 15.9994) / 12.011 = 0.18 kg CO₂

[17] For dairy cows: The carbon loss in the housing units is 2.4 kg (table A.8).
The CH₄ emission is 1.68 kg (see table A.9), which corresponds to 1.68 kg * 12.011/(12.011 + 4*1.008) = 1.25 kg carbon.
Carbon loss as CO₂ = 2.4 kg – 1.25 kg = 1.15 kg C, which corresponds to 1.15 kg * (12.011 + 2 * 15.9994) / 12.011 = 4.21 kg CO₂

[18] 0.06-0.07 litres diesel per 1 m³ slurry * 42.7 MJ/kg * 0.82 kg/litre / 3.6 MJ/kWh / 1.053 kg slurry per m³  = 0.55-0.65 kWh per 1000 kg slurry.

[19] For pig slurry: 0.138 g NH₃-N per g NH₄⁺-N in the slurry “ex storage” * 0.75 kg NH₄⁺-N per kg N * 4.80 kg N = 0.50 kg

[20] For cattle slurry: 0.217 g NH₃-N per g NH₄⁺-N in the slurry “ex storage” * 0.60 kg NH₄⁺-N per kg N * 5.79 kg N = 0.75 kg

[21] The estimate is based on the data in Poulsen and Rubæk (2005, page 5 and page 14): The average leaching of P from agriculture is (690 + 1300 tons P)/2 = 995 tons P per year. The input of P from agriculture to fields is 55000 tons P from manure + 17300 tons P from mineral fertilisers + 5800 tons from sewage sludge and atmosphere = 78100 tons P. Leaching of P is then estimated to 1.2% of the total amount of P applied to the fields (i.e. 955 tons P / 78100 tons P input).

[22] In Danish: Dyre-enhed (DE)

[23] Poulsen and Rubæk (2005, page 24) assumes that if P in the feed is reduced, it will lead to a reduced application of P in animal manure, leading to an overall reduction “... assuming that a reduction in mineral feed phosphates is not counteracted by an increase in the use of mineral fertilizer phosphate” and Poulsen and Rubæk (2005, page 159) “Handelsgødningsfosfor tildeles typisk årligt i mængder, der er afpasset efter jordens fosforstatus og afgrødens behov”

[24] In Danish: Dyre-enhed, DE

[25] 1 kg P = (2 * 30.97376 g/mol + 5 * 15.9994 g/mol) / (2* 30.97376 g/mol ) kg P2O5 = 2.291 kg P2O5

[26] 1 kg K corresponds to (2*39.0983 g/mol + 15.9994 g/mol)/ (2 * 39.0983 g/mol) = 1.205 kg K2O.

| Front page | | Contents | | Previous | | Next | | Top |

Version 1.0 July 2009, © Danish Environmental Protection Agency