Mitigating the environmental impacts of milk production via anaerobic digestion of manure: Case study of a dairy farm in the Po Valley

Science of the Total Environment 481 (2014) 196–208

Contents lists available at ScienceDirect

Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv

Mitigating the environmental impacts of milk production via anaerobic digestion of manure: Case study of a dairy farm in the Po Valley F. Battini a,⁎, A. Agostini b,c, A.K. Boulamanti b, J. Giuntoli b, S. Amaducci a a b c

Institute of Agronomy, Genetics and Field Crops, Università Cattolica del Sacro Cuore, Via Emilia Parmense 84, 29122 Piacenza, Italy STU Unit, JRC-IET—European Commission, Westerduinweg 3, 1755LE Petten, The Netherlands ENEA—Italian National Agency for New Technologies, Energy and the Environment, Via Anguillarese 301, Rome, Italy

H I G H L I G H T S • • • • •

Biogas from manure is a valid option for GHG emission mitigation. GHG emissions of an intensive dairy farm in Northern Italy amount to 1.21 kg CO2 eq. kg− 1 FPCM. If manure is digested in a biogas plant, GHG emissions decrease by 23.7 % if the digestate is stored in an open tank. If manure is digested in a biogas plant, GHG emissions decrease by 36.5 % if the digestate is stored in a gas tight tank. Manure digestion in a biogas plant significantly influences other local environmental impacts.

a r t i c l e

i n f o

Article history: Received 26 September 2013 Received in revised form 4 February 2014 Accepted 9 February 2014 Available online 2 March 2014 Keywords: Life cycle assessment Dairy farm Biogas Greenhouse gas Environmental impacts

⁎ Corresponding author. Tel.: +39 05235992236. E-mail address: [email protected] (F. Battini).

http://dx.doi.org/10.1016/j.scitotenv.2014.02.038 0048-9697/© 2014 Elsevier B.V. All rights reserved.

a b s t r a c t This work analyzes the environmental impacts of milk production in an intensive dairy farm situated in the Northern Italy region of the Po Valley. Three manure management scenarios are compared: in Scenario 1 the animal slurry is stored in an open tank and then used as fertilizer. In scenario 2 the manure is processed in an anaerobic digestion plant and the biogas produced is combusted in an internal combustion engine to produce heat (required by the digester) and electricity (exported). Scenario 3 is similar to scenario 2 but the digestate is stored in a gas-tight tank. In scenario 1 the GHG emissions are estimated to be equal to 1.21 kg CO2 eq. kg−1 Fat and Protein Corrected Milk (FPCM) without allocation of the environmental burden to the by-product meat. With mass allocation, the GHG emissions associated to the milk are reduced to 1.18 kg CO2 eq. kg−1 FPCM. Using an economic allocation approach the GHG emissions allocated to the milk are 1.13 kg CO2 eq. kg−1 FPCM. In scenarios 2 and 3, without allocation, the GHG emissions are reduced respectively to 0.92 (−23.7%) and 0.77 (−36.5%) kg CO2 eq. kg−1 FPCM. If land use change due to soybean production is accounted for, an additional emission of 0.53 kg CO2 eq. should be added, raising the GHG emissions to 1.74, 1.45 and 1.30 kg CO2 eq kg−1 FPCM in scenarios 1, 2 and 3, respectively. Primary energy from non-renewable resources decreases by 36.2% and 40.6% in scenarios 2 and 3, respectively, with the valorization of the manure in the biogas plant. The other environmental impact mitigated is marine eutrophication that decreases by 8.1% in both scenarios 2 and 3, mostly because of the lower field emissions. There is, however, a trade-off between non-renewable energy and GHG savings and other environmental impacts: acidification (+6.1% and +5.5% in scenarios 2 and 3, respectively), particulate matter emissions (+1.4% and +0.7%) and photochemical ozone formation potential (+41.6% and +42.3%) increase with the adoption of a biogas plant. The cause of the increase is mostly emissions from the CHP engine. These impacts can be tackled by improving biogas combustion technologies to reduce methane and NOx emissions. Freshwater eutrophication slightly increases (+0.8% in both scenarios 2 and 3) because of the additional infrastructures needed. In conclusion, on-farm manure anaerobic digestion with the production of electricity is an effective technology to significantly reduce global environmental impacts of dairy farms (GHG emissions and non-renewable energy consumption), however local impacts may increase as a consequence (especially photochemical ozone formation). © 2014 Elsevier B.V. All rights reserved.

F. Battini et al. / Science of the Total Environment 481 (2014) 196–208

1. Introduction Livestock activities have significant impacts on all aspects of the environment. Such impacts are increasing and changing rapidly. The Food and Agriculture Organization (FAO) has developed and applied a methodology based on the life cycle assessment (LCA) approach applicable to the global dairy sector (FAO, 2010). According to their results, the global dairy sector contributes 4.0% to the total global anthropogenic GHG emissions. This figure decreases to 2.7% if meat production is excluded (FAO, 2010). Concerning other environmental impacts, Tukker et al. (2006) have found that dairy products are responsible for about 10% of the total anthropogenic eutrophication potential, 6% of the acidification potential and 4% of the photochemical oxidant formation potential due to all products consumed in EU. Hagemann et al. (2011) have reviewed the GHG emissions of bovine milk production systems in 38 countries and reported that GHG emissions range between 0.8 and 3.07 kg CO2 eq. kg−1 milk. They concluded also that enteric and manure related emissions accounted for 70–95% of the total dairy farm GHG emissions. Nguyen et al. (2013) have assessed several combinations of dairy cattle breeds and feed types in terms of environmental performances and found that enteric fermentation emissions provided the highest contribution to the climate change impact category (45–50%). Their results range between 0.85 and 1.62 kg CO2eq. kg−1. Fantin et al. (2012) have reported GHG emissions ranging between 0.8 and 1.5 kg CO2 eq. kg−1 milk for a collection of European LCA studies. The variations within the range are due not only to the different environmental conditions and farming systems, but also to the assumptions and models used in each study. Del Prado et al. (2013) have modeled a dairy farm in Northern Spain and found GHG emissions of 0.84–2.07 kg CO2 eq. kg−l milk. They also provided evidence that the methodology choice used for the assessment had a large effect on the results. Moreover, they concluded that methane from the rumen and manures, and N2O emissions from soils, were responsible for most of the GHG emissions for milk production. Kristensen et al. (2011) have analyzed 35 conventional farms and 32 organic farms and found global warming emissions, before allocation, of 1.2 and 1.27 kg CO2 eq. kg−l ECM (Energy Corrected Milk), respectively. They developed a new method for the allocation to milk and meat and compared it to 4 others already in use, finding that the share of emissions allocated to milk may vary from 74% to 87%. They identified farming strategies based on low stocking rate or with focus on high efficiency in the herd as the most promising for reducing GHG emissions. Yan et al. (2011) analyzed thirteen LCA studies of European milk production and found that technical issues, arbitrary choices and assumptions make direct comparison between studies challenging. According to Weiske et al. (2006), mitigating the impacts derived from dairy farms is possible by means of the following techniques: (1) improved efficiency of dairy cows; (2) frequent removal of manure and use of scraping systems; (3) improved manure management; and (4) biogas production by anaerobic digestion (AD). Gerber et al. (2011) have analyzed the relationship between productivity of dairy cows and GHG emissions per kg FPCM on global scale and found that GHG emissions decline substantially as animal productivity increases. Anaerobic digestion of manure to biogas is an interesting option because it reduces firstly direct emissions from slurry storage, and, secondly, the emissions from the fossil system replaced (Maranon et al., 2011). Boulamanti et al. (2013) analyzed the environmental impact of several biogas to electricity scenarios, in order to evaluate the sustainability of this process. Their scenarios included maize, manure and codigestion of the two. They found that when using manure, GHG savings higher than 100% were possible and showed that one of the most crucial factors is the management of the digestate. In fact, while the storage of slurry in gas-tight tanks is not generally promoted by agricultural policies, energy and climate policies are

197

starting to subsidize best practices for digestate storage (MSE, 2012). According to Holm-Nielsen et al. (2009), at least 25% of all bioenergy in the future can originate from biogas produced from wet organic materials which include animal manure. In the 27 European Union countries (EU-27) the amount of manure produced is about 1500 million tonnes per year. Manure storage is the second largest source of methane emissions (after enteric fermentation) from European farms (Sneath et al., 2006). Statistics are not easily available but a recent survey indicates that only about 50.7 million tonnes (equal to 8.5% of all slurry produced in EU-27) was treated via AD in EU in 2011 (Lyngsø Foged et al., 2011). In Italy the production of biogas by AD is subsidized by the government (MSE, 2012) and a large increase in the number of farm biogas plants has been recorded in the last years (Fabbri et al., 2013). In most cases, these plants, especially the small ones, apply uncovered storage of digestate. The aim of this study is to carry out an LCA on a representative dairy cattle farm in North Italy. The data were collected from a real farm in the Lodi district. The environmental impacts from the milk production processes are calculated and an assessment of the effects of introducing a biogas plant is carried out. 2. Materials and methods Life cycle assessment (LCA) is a structured, comprehensive and internationally standardized method that aims at quantifying all relevant emissions and resources consumed and the related environmental and health impacts and resource depletion issues that are associated with any product or service. LCA is widely acknowledged as the most suitable tool to assess the environmental impacts of a product or a process (ISO, 2006a; IES, 2010). However, LCA has some inherent sources of uncertainty linked to: the exogenous data used to model the background system (normally from commercial databases); the unavoidable assumptions and the approach used to model the system under analysis (Basset-Mens et al., 2009). Uncertainty is particularly relevant in the case of agricultural systems because of the great variability in the farming practices and the local soil and climate conditions (Flysjö et al., 2011). Comparing results of different LCA studies may be misleading as estimates may vary significantly depending on the assumptions, models and data used (Flysjö et al., 2012). However, comparing systems within the same study (with the same assumptions, input data and models), may lead to reliable conclusions that can be used to provide scientifically sound support to policy makers. This LCA is performed according to the ISO 14040 and 14044 standards (ISO, 2006a, 2006b), using the software GaBi 6 from PE International (PE International, 2013). In Section 2.1 the goal and the scope of the analysis are defined. In Section 2.2 the input data are presented (the life cycle inventory (LCI)). In Section 3 the emissions and resource consumption derived from the LCI are assigned to each impact category analyzed and aggregated into indicators using weighting factors. The results obtained are then interpreted and discussed. The conclusions from the study are reported in Section 4. 2.1. Goal and scope definition The goal of this study is the analysis of the changes in the environmental impacts of a typical Northern Italian dairy farm due to the adoption of a biogas plant running on manure. The following scenarios are compared: Scenario 1: dairy farm without biogas plant Scenario 2: dairy farm with biogas plant and AD of manure with open storage of digestate Scenario 3: dairy farm with biogas plant and AD of manure with covered storage of digestate.

198

F. Battini et al. / Science of the Total Environment 481 (2014) 196–208

The LCA is of the attributional comparative type, the intended audience is the scientific community, policymakers and stakeholders (IES, 2010). 2.1.1. Definition and characteristics of the typical dairy farm in Northern Italy The farm selected in the present study does not represent an average farm in northern Italy, but it is representative of a medium to an efficient, large size scale farm (Guerci et al., 2012), suitable for the development of a biogas plant. The agricultural area of the selected farm is 75 ha; the whole agricultural area is cultivated for the production of cattle feed. The crops cultivated are maize, ryegrass, lucerne (alfalfa) and permanent grassland (Table 1). The soil of the farm is Gleyc Luvisols (loam medium texture with normal fertility). On the farm there are on average 185 Italian Friesian dairy cows, with an average weight of 650 kg. On average, there are 151 lactating cows, 34 dry cows, 85 heifers older than 12 months, 50 heifers younger than 12 months and 30 suckling calves. The internal replacement rate is 31%. The average feed composition of daily lactating cows consists of 18 kg maize silage, 8 kg hay silage, 1.8 kg hay and/or straw and 14 kg concentrates. Feed is distributed as a total mixed ration. Dairy cows are housed in cubicles with polyethylene foam mattresses. A concrete floor with scraping system is used in the cows feeding alley and in the feeding and resting area for heifers. No bedding materials (e.g. straw) are used. Total livestock units (LU) are 279 and the density is equal to 3.72 LU ha−1. The annual milk and meat production of the farm is shown in Table 1. The average milk content is 3.5% protein and 4% fat. The animals sold in a year are 52 dairy cows, 11 heifers and 99 calves. 2.1.2. Functional unit and allocation The functional unit of the analysis is 1 kg of fat and protein corrected milk (FCPM) at the farm gate. FPCM is calculated with the formula defined by the International Dairy Federation (IDF, 2010) (Eq. (1)).     ‐1 −1 ¼ Production kg yr FPCM kg yr

ð1Þ

½0:1226 Fat% þ 0:0776 True Protein% þ 0:2534

In the present study milk is considered as the main product while co-products are meat (all scenarios) and energy (scenarios 2 and 3). In scenario 1, where the only co-products are the animals sold for meat, manure is totally used as organic fertilizer on the farm and the

Table 1 Farm area utilization, yields, production and prices. Actual data are collected from farm operations in 2010–2011. Source: CLAL (2013). ha−1)

Utilization of forage

Crop

Area (ha)

Yields (t

Ryegrass Maize Maize after ryegrass Lucerne Permanent grassland

26 35 26 8 6

15 11.71 53.84 32.5 15

Production

Amount

Unit

Milk Meat

1799.9 48.1

Tonnes FPCM Tonnes live weight

Prices a

Amount

Unit

Milk price Meat price

0.35 0.94

€/kg €/kg

a

F.M.

Prices are an average of 5 years from 2007 to 2011.

Silage Grain Silage Round bale silage Hay

crops are exclusively produced to feed the cattle. Three different types of allocations are applied in this study: • No allocation: all the impact is allocated to the main products, in this case the milk. • Mass allocation based on the weight of the products leaving the farm: the impact is shared between the milk (97.4%) and the animals sold (2.6%). • Economic allocation based on the value of the products sold: in this case the allocated share of the impact is equal to 93.4% to the milk and to 6.6% to the animals sold. For the milk, the average price of farm-gate raw milk in Lombardy over the years 2007–2011 is considered (CLAL, 2013). For the animals sold, the actual price realized by the farm in the same years is used. No subsidies are considered in this study. In scenarios 2 and 3 the additional co-product is electricity. For these scenarios, electricity is not considered directly as a product in the allocation, but indirectly as a credit derived from the avoided emissions in the production of electricity from the Italian grid mix (PE International, 2013). The digestate is totally used as fertilizer in the farm. Slurry or digestate is spread with trailing hoses. No credits for mineral fertilizers replacement are given to the digestate because the fertilizing properties of digestate and slurry are considered equivalent in the long term (Chantigny et al., 2007; Loria and Sawyer, 2005; Möller et al., 2008). 2.1.3. System boundaries The study is from “cradle-to-farm gate”. Fig. 1 summarizes the system boundaries for the different scenarios. The pathway can be divided into the following steps: forage production, milk production and energy production. Disposal and treatment of wastes are included in the analysis. Vitamin supplements, medicines and bull semen are not included; their impact is, though, estimated to be negligible. 2.2. Life cycle inventory (LCI) Real data from the selected farm were collected for a period of two years, 2010 and 2011. To verify the reliability of the data collected they were compared with the average Italian data derived from literature (Guerci et al., 2012; Rossi and Gastaldo, 2012; Zappavigna, 2010). The datasets from actual operations on the farm are complemented by datasets from commercial databases (Ecoinvent, 2010; PE International, 2013) for background processes. 2.2.1. Forage production The forage production inputs are the raw materials (mineral fertilizers, pesticides, herbicides, seeds, plastic wraps and twines, water) and infrastructure (machinery and their shed) necessary for the cultivation of forage crops and grassland (Table 2). Actual consumption data of diesel and lubricating oil for agricultural machinery and transportation are reported in Table 2 for forage and milk production combined. The variability of local climate conditions, soil properties, timing of fertilization and agricultural practices of fertilization and irrigation generate a substantial, complex and difficult to estimate difference in the respective fertilizing properties of manure and digestate. The amount and availability of P and K and micronutrients are not affected by the digestion process as they are not changed in their chemical form and neither volatilized nor lost. A small amount of N is lost during the digestion, storage and spreading of the digestate in comparison to the slurry (as explained in Section 3.6). However, the share of NH4-N has been shown to increase with the digestion, thus, in the short term, the fertilizing effect of digestate is normally higher than that of manure (Schröder and Uenk, 2006; Birkmose, 2007), but the higher N fertilizer value in the year of application is offset by a lower residual N effects in the three next years (Schröder and Uenk, 2006). As a result, it is assumed in this work that the digestate simply replaces manure as a fertilizer. Further research will be needed to

F. Battini et al. / Science of the Total Environment 481 (2014) 196–208

Scenario 1

199

Scenario 2 / 3

Fig. 1. Schematic flowchart of the “cradle-to-farm gate” system (a) without biogas plant (scenario 1) and (b) with biogas plant (scenarios 2 and 3).

estimate the long-term effects of digestate fertilization on other soil properties such as carbon content. The field emissions (Table 3) are mainly due to the fertilization. N fertilization consists of the application of slurry (or digestate) and urea and is responsible for the emissions of nitrous oxides, nitrogen oxides and ammonia. Data from several literature sources were used in order to calculate these emissions and the details are shown in Table 3. Indirect nitrous oxide emissions are calculated according to IPCC (2006) from ammonia and nitrogen emissions to air and from nitrate leaching (for manure management, diesel combustion and CHP emissions). Nitrate leaching is related to N surplus, hydrological conditions, soil type and agronomic management (fertilization and irrigation). In this work, nitrate leaching is estimated from the nitrogen surplus, using the model implemented by Perego et al. (2012) in the Po Valley. The N surplus is calculated by means of a simplified balance: the inputs are from slurry, mineral fertilizer, nitrogen fixation by lucerne and atmospheric deposition; the outputs are the crop removals. In this study the phosphate run-off (including soil erosion) is estimated to be 0.01 kg P per kg P from fertilizer, according to Rossier (1998) and van der Werf et al. (2009). This estimate includes leaching, that is however negligible given the soil characteristics.

2.2.2. Milk production The second step of the pathway consists in the milk production. This process is the same for all three scenarios. The inputs are the raw materials (detergents, paper and water), energy carriers (natural gas and electricity), infrastructures and purchased feed (Table 4). The products of this step are milk, meat and manure. The farm annual production of manure is 6950 m3 containing 80 g kg−1 volatile solids (VS), 4.0 g kg−1 N and 0.8 g kg−1 P. This manure production is sufficient to feed a biogas plant of 50 kWe electricity generating capacity. The emissions due to this step can be grouped as: • enteric and animal house emissions (animal emissions) • emissions from manure storage (in scenario 1) or emissions from digestate storage (in scenario 2) (storage emissions) • emissions from energy carriers (electricity and natural gas). Emissions associated with direct land use change for the purchased feed are not considered in the base case but their influence is analyzed in Section 3.2.2. The enteric emissions of methane are calculated according to the equations reported by Ellis et al. 2007, as reported in Table 3 while

Table 2 Forage production inputs for the whole farm (75 ha). Data are collected from the actual farm operations between the years 2010–2011. Elements

Unit −1

Plastic wraps and twines kg yr Tires kg yr−1 Fertilizers as N before spreading t N yr−1

Maize seeds Grass seeds Herbicides Insecticides Machinery Machinery shed Operations by contractors Irrigation water Diesel Lubricating oil Lorry transport a

a.s. = active substance.

kg yr−1 kg yr−1 kg yr−1 a.s.a kg yr−1 a.s.a kg yr−1 m2 yr−1 m2 yr−1 m3 yr−1 kg yr−1 kg yr−1 t km yr−1

Quantity

Comments

311 250 26.9 25.4 6.81 1598 1091 138 95 6667 12.5 874,400 84,000 23,372 120 74,732

Included disposal and treatment of waste Included disposal and treatment of waste Slurry Digestate Urea, used only for maize 180 units of 25,000 seed Lucerne seed and ryegrass seed Only for maize (S-metoalaclor, Terbutilazina) Only for maize (against soil insects and Ostrinia nubilalis) Average lifetime 12 years Average lifetime 40 years Harvest of maize grain and silage of grass and maize Flood and spray irrigation with superficial water (three irrigation periods for the maize and grassland; one for Lucerne) For agricultural machinery (for forage and milk production together) Included disposal and treatment of waste (for forage and milk production together) For purchased raw materials (for forage and milk production together)

f

e

d

c

Nitrous oxide direct emissions occurring after field application of the slurry are calculated according to IPCC (2006), while the ones occurring after urea application are estimated on the basis of the data collected by Sanz-Cobena et al. (2012). Nitrogen oxides emissions from slurry are estimated according to Stehfest and Bouwman (2006) and from urea according to Sanz-Cobena et al. (2012). Ammonia emissions are estimated according to Amon et al. (2006) for slurry and according to EMEP/EEA (2009) for urea. Emissions from diesel tractors used on farm. Consumptions and emissions for feedstocks transport are taken from Ecoinvent (Ecoinvent v2.2, 2010). Calculated as 2% of scenario 2 storage emissions. Specific enteric emissions are equal to: 145.3 kg CH4/year for lactating cows, 86 kg CH4/year per dry cows, and 64 kg CH4/year per heifers. b

kg/yr kg/yr kg/yr kg/yr kg N/yr kg P/yr SO2 CO NMVOC TSP N leaching P run-off

13,272 55.6

2769c 2531c kg/yr NH3

11,859 55.6

Perego et al. (2012) Rossier (1998) and van der Werf et al. (2009)

Winther and Nielsen (2006) Winther and Nielsen (2006) Winther and Nielsen (2006) Winther and Nielsen (2006) 23.61 479.2 100.9 74.3

429b 447b kg/yr NOx

a

Amon et al. (2006) 1.38 69 EMEP/EEA(2009) Winther and Nielsen (2006) 0.15

Winther and Nielsen (2006) 885.2

3211

11,090 33469f Winther and Nielsen (2006) Winther and Nielsen (2006) Winther and Nielsen (2006) 74499.9 1.7 9.6

Amon et al. (2006) IPCC (2006) and Sanz-Cobena et al. (2012) Stehfest and Bouwman (2006) and SanzCobena et al. (2012) Amon et al. (2006) and EMEP/EEA (2009) 14 745a kg/yr kg/yr kg/yr CO2 CH4 N2O

9 783a

Source Scenario 2/3 Scenario 1

Unit

Field emissions

285

IPCC (2006) 1.1 55 59

187 3.98 9331 199 Stoichiometric calculations Ellis et al.(2007)

28,118 145

Scenario 2 Scenario 1

Source Diesel emissionsd

Animal emissions

Source

Storage emissions

Scenario 3e

Source

Amon et al. (2006) Amon et al. (2006)

F. Battini et al. / Science of the Total Environment 481 (2014) 196–208

Emission

Table 3 Emissions from various activities of the whole farm: field emissions, emissions from the combustion of diesel for agricultural operations, animal emissions (enteric and animal housing) and finally emissions from manure and digestate storage.

200

nitrous oxide emissions from animal houses that do not use beddings are very low, and can be neglected (Zhang et al., 2005). The annual enteric emissions are equal to 145.3 kg CH4/year per lactating cow. The ammonia emissions are quantified according to the international guidelines (EMEP/EEA, 2009) as reported in Table 3. Stoichiometric calculation of CO2 emissions from calcium carbonate used as mineral supplement in the feed is also included (Table 3). Emissions from the storage of manure (Table 3) depend on several factors: manure type, storage temperature, storage duration, aeration, and formation of a natural crust at the surface of slurry, to name the most important ones. These emissions were reported in several studies (Massé et al., 2003; Sneath et al., 2006; Dinuccio et al., 2011; Chadwick et al., 2011). In most farms in Northern Italy, the cattle are reared in the stable all year round (no grazing) and the manure produced is stored in open tanks before it is spread on the fields. This is the case also for the reference scenario (scenario 1). The storage of slurry takes place under anaerobic conditions, generally without any treatment. Normally a surface crust is formed, which leads not only to a decrease of methane and ammonia emissions, but also to an increase of nitrous oxide emission (Smith et al., 2007; Sonesson et al., 2009). In this study, the storage emissions of methane, (Table 4) are estimated according to Amon et al. (2006). The slurry N content is the actual data measured at the farm (see Section 2.2.2). For slurry management, the N losses were calculated according to Amon et al. (2006) for NH3 and N2O; according to IPCC (2006) for NOx and according to Oenema et al. (2000) for N2. The losses due to spreading and field emissions are calculated according to Amon et al. (2006) for NH3, according to IPCC (2006) for N2O and according to Stehfest and Bouwman (2006) for NOx. The losses due to spreading and field emissions of urea are calculated according to EMEP/CORINAIR (2007) for NH3 and Sanz-Cobena et al. (2012) for N2O and NOx. N losses due to crop residues decay are calculated according IPCC (2006). The model used to calculate the N leaching is from Perego et al. (2012). In scenarios 2 and 3 additional losses come from the digestion process (about 6% of the N content in slurry (Schievano et al., 2011)). All the other losses are calculated with the same models used in scenario 1 but applied to the digestate rather than to the slurry. From the stable to the field, the total N losses in scenario 1 are calculated to be equal to 2.3% while in scenarios 2 and 3 they amount to 7.8%. Because of the relevance and high variability of manure and digestate storage emissions, a sensitivity analysis is carried out to assess the impact of a 30% storage emissions increase or decrease on global warming. The results are reported in Section 3.2.1. Emissions from the supply of electricity and natural gas are taken from PE International for the reference year 2010 (PE International, 2013). 2.2.3. Energy production This step is present only in scenarios 2 and 3 (Fig. 1). The two phase biogas plant consists of two mixed anaerobic digesters, operating under mesophilic conditions (about 38 °C) and of two open storage tanks for scenario 2 or gas-tight storage tanks with double membrane sheet for scenario 3. The biogas produced is firstly desulfurized and dehumidified and then converted to electrical energy by means of a reciprocating internal combustion engine combined heat and power unit (CHP). The main characteristics of the biogas plant are shown in Table 5 and are based on the data by Fabbri et al. (2011) for a biogas plant operating with only cattle slurry. The operating time is about 8000 h per year. The energy production inputs are infrastructures and lubricating oil; the datasets are taken from Ecoinvent (Ecoinvent, 2010). Although it is called a CHP engine, only electricity is considered as output in this study, since the heat produced is not exported. From the thermal energy produced only a part is used to supply the process heat required by the digester. The electrical energy is delivered to the

F. Battini et al. / Science of the Total Environment 481 (2014) 196–208

201

Table 4 Milk production inputs for the whole farm. Data are collected from the actual farm operations in 2010–2011. Elements

Unit

Quantity

Comments

Detergents Paper Tap water Natural gas Electricity Loose housing system Hay from grassland Straw Soybean meal Cotton seed Pulps from sugar beet Molasses from sugar beet Substituted powder milk Mineral supplements Trans-oceanic transport

kg yr−1 kg yr−1 m3 yr−1 MJ yr−1 MJ yr−1 unit yr−1 kg yr−1 kg yr−1 kg yr−1 kg yr−1 kg yr−1 kg yr−1 kg yr−1 kg yr−1 t km yr−1

2100 100 10,184 221,112 293,897 0.20 30,000 140,000 372,866 58,000 9928 69,000 2800 38,913 5,601,700

Alkaline and acid detergents Included disposal From well

grid (329.77 MWh yr−1 in scenario 2 and 368.79 MWh yr−1 in scenario 3). The higher production in scenario 3 is due to the recovery of additional biogas from the digestate storage. As mentioned earlier, the production of electricity is considered as a credit. The credit is estimated on the basis of saved emissions from the electricity production from the Italian electricity mix. CHP emissions are estimated according to Kristensen and Jensen (2001) for biogas fired CHP units, the emission factors are shown in Table 5. A leak of 1% of the methane produced in the biogas plant is assumed to account for the accidental emissions due to membrane cover permeability (Zifu Li et al., 2013), leaky gaskets, maintenance operations and flaring or venting of the overproduction (Liebetrau et al., 2010) for both scenarios 2 and 3. In scenario 3 the emissions from digestate storage, mainly due to the handling of the digestate, are assumed as 2% of the digestate storage emissions in scenario 2.

3. Results and discussions 3.1. Life cycle impact assessment (LCIA) The inventories of emissions and resources consumed are assessed in terms of impacts, in order to understand and evaluate their magnitude and significance. The following environmental impact categories were analyzed: global warming, acidification,

Table 5 Main characteristics of the biogas plant (Fabbri et al., 2011) and emissions from the CHP engine per GJ biogas. Adapted from Kristensen and Jensen (2001)). Parameter Biogas production Methane content in biogas Methane lower calorific value Gross electrical efficiency Power consumption of digester Power consumption of CHPa Volume of digester Retention time Digester life time CHP life time NOx Methane NMVOC Carbon monoxide Nitrous oxideb Formaldehyde Sulfur dioxide a b

Unit 3

Value −1

m kg vs % MJ per Nm3 % % of electricity produced % of electricity produced m3 days years years g GJ−1 g GJ−1 g GJ−1 g GJ−1 g GJ−1 g GJ−1 g GJ−1

According to Fabbri and Baldrighi (2012). Including indirect emissions calculated according to IPCC (2006).

0.4 55 36 32 11.9 3% 600 30–35 20 10 540 323 14 273 3,96 21.15 19

Low voltage at farm Unit for 22 cows and rearing — average lifetime 40 years Purchased feed Purchased feed (not for bedding) Purchased feed Purchased feed Purchased feed Purchased feed Purchased feed Calcium carbonate, magnesium oxide, calcium phosphate, sodium chloride For soybean meal and cotton seeds

particulate matter/respiratory inorganics, photochemical ozone formation, freshwater and marine eutrophication. The assessment is performed at midpoint using the methods recommended by the ILCD Handbook (IES, 2012). The midpoint methods are characterization methods that provide indicators for comparison of environmental interventions at a level of the cause–effect chain between emissions/resource consumption and the endpoint level where the endpoint level is an attribute or aspect of natural environment, human health, or resources, identifying an environmental issue giving cause for concern (IES, 2012). In addition, two technical quantities (primary energy from non-renewable resources and land use) are included in the analysis. The results for all the impact categories analyzed are summarized in Table 6. As mentioned earlier, three allocation methods are presented in this study, however, for the sake of simplicity, the results discussed in the following paragraphs do not consider allocation. In fact, allocating the impacts to the meat co-product would only change proportionally the impacts of all the scenarios without interfering in the conclusions that can be drawn. 3.2. Global warming potential The impact on climate change is assessed using the IPCC model characterisation factors, also known as GWP, at the 100-year horizon (IPCC, 2007). The unit for the characterisation is kg CO2 eq. The total GHG emissions for scenario 1 are calculated to be 1.21 kg CO2 eq. kg− 1 FPCM as shown in Fig. 2a. This value is in line with other cases such as, for example, the ones collected by Fantin et al. (2012) reporting GHG emissions ranging between 0.8 and 1.5 kg CO2 eq. kg−1 FPCM. By introducing a biogas plant with open digestate storage in scenario 2, the total GHG emissions decrease by 23.7, down to 0.92 kg CO2 eq. kg−1 FPCM. This result is due to the avoided slurry storage emissions (which are considerably higher than for digestate storage) and to the credit from the electricity produced (the emissions from Italian electricity production amount to 0.146 kg CO2 eq. MJ−1). In the case of scenario 3 the GHG emissions decrease by 36.5%, down to 0.77 kg CO2 eq. kg−1 FPCM, mainly because, in addition to the avoided slurry storage emissions, digestate storage emissions are captured and the methane contained in them is used for additional energy production. The four main processes contributing to global warming in scenario 1 (Fig. 2b) are: enteric emissions from the animals (34.7%), emissions from slurry storage (30.7%), field emissions (10.7%) and the purchased feed (9.6%, of which around 80% derived from the imported soybean meal). In all scenarios the main contributor to global warming is methane (64%, 62% and 60% in scenario 1, 2 and 3, respectively), derived mostly

202

F. Battini et al. / Science of the Total Environment 481 (2014) 196–208

Table 6 Results for all the impacts considering no allocation, mass and economic allocation (per kg−1 FPCM). No allocation

Mass allocation

Economic allocation

kg CO2 eq. Mole of H+ eq. kg P eq. kg N eq. kg PM 2.5 eq. kg NMVOC eq. MJ m2

1.21 1.31E−02 1.23E−04 9.37E−03 4.15E−04 2.83E−03 3.68 1.00

1.18 1.28E−02 1.20E−04 9.13E−03 4.05E−04 2.76E−03 3.59 0.98

1.13 1.23E−02 1.15E−04 8.75E−03 3.88E−04 2.64E−03 3.44 0.94

Scenario 2 GWP Acidification Eutrophication freshwater Eutrophication marine P.M./respiratory inorganics Photochemical ozone form. Non-renewable energy Land use

kg CO2 eq. Mole of H+ eq. kg P eq. kg N eq. kg PM 2.5 eq. kg NMVOC eq. MJ m2

0.92 1.39E−02 1.24E−04 8.61E−03 4.21E−04 4.01E−03 2.35 1.00

0.90 1.35E−02 1.21E−04 8.39E−03 4.10E−04 3.91E−03 2.29 0.98

0.86 1.30E−02 1.16E−04 8.04E−03 3.93E−04 3.74E−03 2.19 0.94

Scenario 3 GWP Acidification Eutrophication freshwater Eutrophication marine P.M./respiratory inorganics Photochemical ozone form. Non-renewable energy Land use

kg CO2 eq. Mole of H+ eq. kg P eq. kg N eq. kg PM 2.5 eq. kg NMVOC eq. MJ m2

0.77 1.38E−02 1.24E−04 8.61E−03 4.18E−04 4.03E−03 2.19 1.00

0.75 1.34E−02 1.21E−04 8.39E−03 4.07E−04 3.92E−03 2.13 0.98

0.72 1.29E−02 1.16E−04 8.04E−03 3.91E−04 3.76E−03 2.04 0.94

Scenario 1 GWP Acidification Eutrophication freshwater Eutrophication marine P.M./respiratory inorganics Photochemical ozone form. Non-renewable energy Land use

from enteric emissions. Electricity credits have a significant impact on fossil-CO2 emissions that are reduced by 37% and 42% in scenarios 2 and 3, respectively. The introduction of a biogas plant to the farm does not substantially change the results of the contribution analysis because many processes and inputs do not change their absolute amounts (e.g. purchased feed, diesel, electricity, fertilizer production and enteric emissions). A small increase of emissions from infrastructure occurs, due to the construction of the biogas plant, but this is still very limited (2.3% in scenario 2 and 2.7% in scenario 3). Additional emissions from end-use combustion of the biogas are included, but they account only for about 3.5% and 4.2% of the total impact in scenarios 2 and 3, respectively. Fig. 2b clearly shows that the main advantages in the configuration modeled in scenarios 2 and 3 derive from the avoided storage emissions and the credits for substituted fossil electricity. In fact, storage emissions decrease compared to scenario 1 by 60% and by 99% in scenario 3. Field emissions are also slightly decreased (− 5%) because of the lower nitrous oxide emissions from digestate compared to slurry. The credits due to the electricity from biogas are calculated to be equal to 0.096 kg CO2 eq. kg− 1 FPCM in scenario 2 and 0.107 kg CO2 eq. kg−1 FPCM in scenario 3. The reduction of emissions from the electricity production accounts for 33.5% of the savings in the scenario with open digestate storage and 24.3% in the scenario with covered storage. 3.2.1. Sensitivity analysis on storage emissions After enteric emissions, storage emissions are the main contributor to the total GHG emissions, but they are very variable and rarely measured experimentally (IEA Bioenergy, 2013). In order to verify the influence of the assumption made in this study, Fig. 2c shows the results of a sensitivity analysis. The storage emissions are varied by +/−30% both for manure and for digestate storage. This value accounts for the difference between the data used in this study and the values indicated by the IPCC guidelines for similar conditions (IPCC, 2006) and the range reported in the IEA biogas handbook (IEA bioenergy, 2013). Increasing the emissions, the total GHG emissions become 1.32 kg CO2 eq. kg−1 FPCM in scenario 1, 0.97 kg CO2 eq. kg−1 FPCM in scenario 2, and 0.77 kg CO2 eq. kg−1 FPCM in scenario 3. Decreasing

the emissions, instead, the total impact becomes 1.10 kg CO2 eq. kg−1 FPCM in scenario 1, 0.88 kg CO2 eq. kg− 1 FPCM in scenario 2, and 0.77 kg CO2 eq. kg−1 FPCM in scenario 3. Storage emissions for scenario 3, where the emissions are captured, are practically negligible. This sensitivity analysis shows that basically the range of results found in many studies could be simply explained by the variation of this parameter. It is evident that in order to have a precise picture of the GHG emissions from the dairy industry, additional experimental results quantifying storage emissions from manure and digestate management are essential. 3.2.2. Sensitivity analysis on soybean cake related land use change (LUC) emissions All the farm inputs except soybean meal have no or negligible emissions from LUC. This is because they come from processes which have either very limited land use or come from countries where emissions from LUC are very limited (e.g. EU and North America). The Italian domestic production of soya is limited, 468 ktonnes in 2009 (FAOSTAT, 2013), equivalent to about 375 ktonnes of meal. In comparison, soybean meal imports to Italy amounted to 2157 ktonnes in 2009 (FAOSTAT, 2013). The global market of soybean meal is dominated by countries with high LUC emissions such as Brazil and Argentina (Opio et al., 2013). FAO calculated that for each kg of soybean meal internationally traded, 2.98 kg of CO2 eq is emitted because of LUC (FAO, 2010). As each kg of FPCM needs 0.182 kg of soybean meal as input, in each scenario, without allocation, the emissions including LUC due to soybean meal production would significantly increase the GHG emissions by 0.53 kg CO2 eq. kg−1 FPCM (1.74, 1.45 and 1.30 kg CO2 eq. kg−1 FPCM in scenarios 1, 2 and 3, respectively). 3.3. Acidification potential The acidification potential is expressed in moles of H+ eq. kg− 1 FPCM and is calculated according to the accumulated exceedance method (Seppälä et al., 2006; Posch et al., 2008).

F. Battini et al. / Science of the Total Environment 481 (2014) 196–208

203

(a)

(b)

(c)

(d)

Fig. 2. (a) GHG emissions of all the scenarios calculated using the GWP(100) per kg of FPCM. The total emissions and the contribution shares are indicated. (b) GHG emissions from all the scenarios, contribution of the main processes. (c) Results of the sensitivity analysis on storage emissions GHG for scenarios 1 and 2. (d) Non-renewable energy use calculated for the whole farm (75 ha) expressed in MJ.

As shown in Fig. 3a, in all the scenarios the emissions of ammonia are the largest contributor accounting for 82% of the total impact in scenario 1 and 77% in scenarios 2 and 3. The remaining contribution is due to NOx and SO2 emissions. NOx emissions increase with the combustion of biogas in scenarios 2 and 3 while SO2 emissions decrease because of the fossil electricity substituted. The contribution of combustion emissions causes the acidification potential to increase slightly in scenarios 2 and 3, by 6.1% and 5.5%, respectively. The main processes contributing to the acidification impact are the emissions from the animals, accounting for about 40% of the impact in all scenarios, followed by field emissions (about 34% of the impact). Field emissions are slightly increased in scenarios 2 and 3 because the emissions of ammonia derived from the spreading of digestate are slightly higher than those from the use of slurry. In scenario 1 slurry application is responsible for 51% of the ammonia field emissions while in scenarios 2 and 3 this share increases to 55% with digestate application. Scenarios 2 and 3 have an additional 7 and 8% contribution from biogas combustion emissions but this is partly compensated by credits and by reduced storage emissions, Fig. 3b. It is important to note that the NOx emissions from biogas combustion are highly variable depending on the engine type, the technology adopted and the operating conditions.

The adoption of technologies for NOx emission abatement may dramatically reduce NOx emissions (Camarillo et al., 2013).

3.4. Particulate matter/respiratory inorganics Particulate matter emissions are expressed in terms of kg PM 2.5 eq. and are calculated according to the intake fraction concept and use the conversion factors defined by the RiskPoll software (Rabl and Spadaro, 2004) and Greco et al. (2007). These factors have been calculated by the ILCD (IES, 2012) for average EU conditions. Fig. 4a shows that emissions of ammonia and NOx are the main contributors to this impact category because of their role in the formation of secondary particulate matter. Direct emissions of particulate matter account for about 26–27% of the total impact depending on the scenario. The addition of the biogas plant causes a slight increase of the impact, of about 1.4% and 0.7% for scenarios 2 and 3, respectively, mostly because of the CHP emissions. The main processes contributing to this impact (Fig. 4b) are once again emissions of ammonia from the animal housing (about 27–29% of the impact) and from fertilizer application (23–24% of the impact). Purchased feed emissions are responsible for about 13% of the impact

204

F. Battini et al. / Science of the Total Environment 481 (2014) 196–208

(a)

(b)

Fig. 3. (a) Acidification potential (AP) per kg FPCM, calculated according to the accumulated exceedance method for all the scenarios; bars are stacked based on the contribution of the most relevant species. (b) AP, contribution of the different processes.

and infrastructures are not negligible, accounting for 9% of the impact. In the case of scenarios 2 and 3, storage emissions decrease while field emissions slightly increase and there are additional CHP emissions. Storage emissions decrease from 11 mg PM 2.5 eq. kg−1 FPCM in scenario 1 to 3 mg PM 2.5 eq. kg−1 FPCM in scenario 2 and negligible emission in scenario 3. Field emissions, however, increase from 97 mg PM 2.5 eq. kg−1 FPCM for scenario 1 to 105 mg PM 2.5 eq. kg−1 FPCM in both scenarios 2 and 3 because the emissions of ammonia from the digestate are higher than those from manure. There are 18 and 19 mg PM 2.5 eq. kg−1 FPCM additionally in the two scenarios 2 and 3 respectively, because of the CHP while the credits from substituted electricity amount to 15 and 16 mg PM 2.5 eq. kg−1 FPCM. In practice, the introduction of the biogas plant shifts the PM emissions from the regional electricity supply to the local CHP plant.

(a)

3.5. Photochemical ozone formation potential (POFP) The photochemical ozone formation potential is expressed in terms of kg NMVOC eq. and is calculated according to the ReCiPe method (Goedkoop et al., 2008). Fig. 5a shows that for this impact, the addition of a biogas plant has a negative effect. The lower emissions of methane from storage and the credits from substituted electricity, in fact, are not enough to compensate the increase in NOx emissions from the combustion of biogas. This impact shows a high increase of about 41.6% in scenario 2 and 42.3% in scenario 3. The main processes contributing to this impact for scenario 1 (Fig. 5b) are diesel combustion on the farm (26%), the purchased feed (24% of the impact) and transport of the purchased feed (20%). The impact in scenarios 2 and 3 is dominated by biogas combustion emissions (34.5% and 37%) which are not compensated by the credits (−4%).

(b)

Fig. 4. (a) Particulate matter emissions (PM) per kg FPCM, calculated according to the intake fraction concept, for all the scenarios and contribution of the most relevant species. (b) PM emissions, contribution of the different processes.

F. Battini et al. / Science of the Total Environment 481 (2014) 196–208

(a)

205

(b)

Fig. 5. (a) Photochemical oxidant formation potential (POFP) per kg FPCM, calculated according to the ReCiPe method, for all the scenarios and contribution of the most relevant species. (b) POFP, contribution of the different processes.

As mentioned before, NOx emissions from biogas engines are variable, however, technological improvements and emissions control systems may help to limit this impact in the case of energy production on-site. The most promising technology is selective catalytic reduction that may reduce the emissions of NOx by more than 80% (Camarillo et al., 2013). Given the specific issues of ozone formation in the Po Valley, (Mircea et al., 2014), the adoption of the best available technologies to abate NOx emissions should be strongly promoted. 3.6. Freshwater and marine eutrophication The impact on eutrophication is divided into two categories depending on the main substances responsible. In freshwater ecosystems phosphorus is the limiting nutrient, therefore only Pcompound emissions are considered for the assessment of freshwater eutrophication and impacts are expressed in terms of kg P eq. In sea waters the limiting factor for plant growth is normally N, hence the recommended method includes only N compounds in the characterization of marine eutrophication (IES, 2012). The contributing substances are nitrate, ammonia and nitrogen oxides are taken into consideration and the impact is expressed in terms of kg N eq. (IES, 2012). Both categories are calculated according to the ReCiPe method (Goedkoop et al., 2008). Fig. 6a shows that the impact on freshwater eutrophication slightly increases when a biogas plant is added (+ 0.8%). The emissions in all scenarios (Fig. 6b) result mainly from the purchased feed, accounting for almost 55% of the impact, and from field emissions (about 25%). In scenarios 2 and 3 the absolute impacts are the same as in scenario 1, except for the infrastructure and the credits for renewable electricity, with emissions due to the construction of the biogas plant being practically the only reason for the differences between the scenarios since the credits are negligible. As shown in Fig. 6c, the potential eutrophication impact on marine ecosystems is decreased, by approximately 8%, in both scenarios including the biogas plant. The main contributor to this impact is nitrogen leaching and run-off derived from cultivation practices. Field emissions account for about 80% of the impact in all the scenarios (Fig. 6d). The decrease in marine eutrophication is basically

due to the lower nitrogen content in the digestate in comparison to the slurry due to the N losses during digestion and digestate management. End-use emissions in scenarios 2 and 3 have a small contribution (about 0.6%) and credits are negligible due to the already low impact of the current electricity. In practice, freshwater eutrophication does not change with the introduction of the biogas plant because the anaerobic digestion does not affect the P flows if not for the additional emissions linked to the additional infrastructure. Marine eutrophication instead is reduced mainly because of the N lost during the digestion process; there is a limited contribution to the reduction from N lost in the digestate storage and spreading. However, the Northern Adriatic sea, where the Po river discharges the nutrients from the leaching and run-off of the Po Valley, is a case of marine ecosystem in which the limiting nutrient is P. In fact, the Northeastern Adriatic continental shelf has been subjected to an increasing pressure due to a high river transport of nitrogen and, currently, only a deep phosphorus deficiency in the river nutrient pools prevents its severe eutrophication (Cozzi et al., 2012). Therefore, in this specific case, also the marine eutrophication impact should be measured based on P emissions. In that case, the result would be very similar to what is seen for the freshwater impact: no significant change with the introduction of the biogas plant. 3.7. Primary energy from non-renewable resources In scenario 1 (Fig. 2d) the process that requires more primary energy from non-renewable resources is the purchased feed (30.5% of the impact), followed by the diesel consumed on the farm and for the transport of purchased feed, accounting for 22% and 16% of the impact, respectively. In scenarios 2 and 3 it is possible to note an increase of energy needed for infrastructure (+11% compared to scenario 1) due to the construction of the biogas plant. The introduction of the biogas plant in scenario 2 saves 36.2% of the total non-renewable energy use in scenario 1, while in scenario 3 the credits amount to 40.6% of scenario 1. Power production from biogas from manure guarantees a reduction in the consumption of fossil fuels, and, very importantly for Italy, improves the security of energy

206

F. Battini et al. / Science of the Total Environment 481 (2014) 196–208

(a)

(b)

(c)

(d)

Fig. 6. Eutrophication potential for all the scenarios per kg FPCM. (a) Freshwater eutrophication potential (FEP). (b) FEP, contribution of the different processes. (c) Marine eutrophication potential (MEP). (d) MEP, contribution of the different processes.

supply, being a domestic source. An additional benefit of manure biogas power is the possibility of storing the biogas produced and generating power that may be used to compensate the increasing need for balancing power due to the energy supply from fluctuating sources, such as photovoltaic or wind power (Hahn et al., 2014). 3.8. Land use Regarding land use (land occupation), the slight increase due to the biogas plant is practically offset by the reduced use of soil due to the credit from the electricity production. In any case, both have a very limited impact in comparison to the amount of land needed for the cultivation of the feeds for milk production. In fact, the land use amounts to 1 m2 kg−1 FPCM yr−1 for all scenarios. If we consider only the amount of land needed for the electricity production (by the biogas plant or the electricity mix) the amount of land occupied by the biogas plant is about half of the land needed by the Italian electricity mix. The farm itself is responsible for about 42% of the total land occupation, the rest is due to external inputs, mainly the purchased feed (56.3%, of which 78.1% due to soybean meal).

4. Conclusions In this study the environmental impacts from cradle to farm gate of milk production in Northern Italy are quantified. Changes of these impacts achieved with the introduction of a biogas production plant with either open or closed digestate storage are assessed. The results show a significant decrease of GHG emissions when anaerobic digestion is introduced, especially for covered (gas-tight) storage of digestate. The mitigation of GHG emissions is due in part to the substitution of fossil electricity, but most of the savings are achieved thanks to the avoided GHG emissions from undigested slurry storage. Also, non-renewable energy consumption shows a marked decrease with the adoption of the biogas plant. Environmental impacts such as eutrophication, particulate matter and land use are only slightly affected by the implementation of the anaerobic digestion process. The downside is the increment of acidification and photochemical oxidant formation. In particular, this last impact shows a considerable increase. However, the increases in acidification and photochemical oxidant formation depend on the NOx emissions from biogas combustion. It is evident therefore that this negative impact can be tackled by improved biogas combustion

F. Battini et al. / Science of the Total Environment 481 (2014) 196–208

technologies and/or the adoption of post combustion selective catalyst reduction systems to reduce NOx emissions in order to avoid trade-offs between local (photochemical ozone) and global pollution (global warming). Acknowledgment The project is funded by LIFE ENV 09 IT 000214 GAS-OFF — “Integrated strategies for GHG mitigation in dairy farm”. “Associazione Provinciale Allevatori Milano Lodi” has provided the data. References Amon B, Kryvoruchvo V, Amon T, Zechmeister-Boltenstern S. Methane, nitrous oxide and ammonia emissions during storage and after application of dairy cattle slurry and influence of slurry treatment. Agric Ecosyst Environ 2006;112:153–62. Basset-Mens C, Kelliher FM, Ledgard S, Cox N. Uncertainty of global warming potential for milk production on a New Zealand farm and implications for decision making. Int J Life Cycle Assess 2009;14:630–8. Birkmose T. Digested manure is a valuable fertilizer. European Biogas workshop and study tour The Future of Biogas in Europe III. 14th–16th of June 2007. University of Southern Denmark, Esbjerg — Denmark. Proceedings; 2007. p. 89–94. Boulamanti AK, Donida Maglio S, Giuntoli J, Agostini A. Influence of different practices on biogas sustainability. Biomass Bioenergy 2013;53:149–61. Camarillo MK, Stringfellow WT, Hanlon JS, Watson KA. Investigation of selective catalytic reduction for control of nitrogen oxides in full-scale dairy energy production. Appl Energy 2013;106:328–36. Chadwick D, Sommer S, Thorman R, Fangueiro D, Cardenas L, Amon B, et al. Manure management: Implications for greenhouse gas emissions. Anim Feed Sci Technol 2011;166–167:514–31. Chantigny MH, Angers DA, Rochette R, Bélanger G, Massé D. Gaseous nitrogen emissions and forage nitrogen uptake on soils fertilized with raw and treated swine manure. J Environ Qual 2007;36:1864–72. CLAL. Consultancy and market research food and diary; 2013 [www.clal.it visited on December 2013]. Cozzi S, Falconi C, Comici C, Cermelj B, Kovac N, Turk V, et al. Recent evolution of river discharges in the Gulf of Trieste and their potential response to climate changes and anthropogenic pressure. Estuar Coast Shelf Sci 2012;115:14–24. Del Prado A, Mas K, Pardo G, Gallejones P. Modelling the interactions between C and N farm balances and GHG emissions from confinement dairy farms in northern Spain. Sci Total Environ 2013;465:156–65. Dinuccio E, Berg W, Balsari P. Effects of mechanical separation on GHG and ammonia emission from cattle slurry under winter conditions. Anim Feed Sci Technol 2011;166–167: 532–8. Ecoinvent. Ecoinvent Database v2.2. Swiss Centre for Life Cycle InventoriesDübendorf, Switzerland: Swiss Centre for Life Cycle Inventories; 2010. Ellis JL, Kebreab E, Odongo NE, McBride BW, Okine EK, France J. Prediction of methane production from dairy and beef cattle. J Dairy Sci 2007;90:3456–67. EMEP/CORINAIR. Atmospheric Emission Inventory Guidebook. Technical report No 16/2007; 2007. EMEP/EEA. Air Pollutant Emission Inventory Guidebook EEA. Technical report No 9/2009; 2009. Fabbri C, Baldrighi C. Più biogas integrando i reflui zootecnici con silomais. L'informatore agrario 2012;5:30–3. Fabbri C, Piccinini S, Verzellesi F. Biogas da soli reflui zootecnici, buona la resa energetica. L'informatore agrario 2011;38:20–4. Fabbri C, Labartino N, Manfredi S, Piccinini S. Biogas, il settore è strutturato e continua a crescere. L'informatore agrario 2013;11:11–6. Fantin V, Buttol P, Pergreffi R, Masoni P. Life cycle assessment of Italian high quality milk production. A comparison with an EPD study. J Clean Prod 2012;28:150–9. FAO. Greenhouse Gas Emissions from the Dairy Sector: A Life Cycle Assessment. Food and Agriculture Organization of the United Nations RomeItaly: Food and Agriculture Organization of the United Nations Rome; 2010. FAOSTAThttp://faostat.fao.org/site/339/default.aspx, 2013. Flysjö A, Henriksson M, Cederberg C, Ledgard S, Englund JE. The impact of various parameters on the carbon footprint of milk production in New Zealand and Sweden. Agr Syst 2011;104:459–69. Flysjö A, Cederberg C, Henriksson M, Ledgard S. The interaction between milk and beef production and emissions from land use change and critical considerations in life cycle assessment and carbon footprint studies of milk. J Clean Prod 2012;28:134–42. Gerber P, Vellinga T, Opio C, Steinfeld H. Productivity gains and greenhouse gas emissions intensity in dairy systems. Livest Sci 2011;139:100–8. Goedkoop MJ, Heijungs R, Huijbregts M, De Schryver A, Struijs J, Van Zelm R, et al. A life cycle impact assessment method which comprises harmonised category indicators at the midpoint and the endpoint level; 2008 [First edition Report I: Characterisation, 6 January 2009, http://www.lcia-recipe.net (last accessed 12 September 2013)]. Greco SL, Wilson AM, Spengler JD, Levy JI. Spatial patterns of mobile source particulate matter emissions-to-exposure relationships across the United States. Atmos Environ 2007;41:1011–25. Guerci M, Zucali M, Bava L, Sandrucci A, Tamburini A. Piu' energia netta latte riducendo l'impatto ambientale. L'informatore agrario 2012;32:29–32.

207

Hagemann M, Hemme T, Ndambi A, Alqaisi O, Sultana N. Benchmarking of greenhouse gas emissions of bovine milk production systems for 38 countries. Anim Feed Sci Technol 2011;166–167:46–58. Hahn H, Krautkremer B, Hartmann K, Wachendorf M. Review of concepts for a demand-driven biogas supply for flexible power generation. Renew Sustain Energy Rev 2014;29:383–93. Holm-Nielsen JB, Al Seadi T, Oleskowicz-Popiel P. The future of anaerobic digestion and biogas utilization. Bioresour Technol 2009;100:5478–84. IEA Bioenergy. The Biogas Handbook: Science, production and application; 2013 [ISBN 0 85709 498 X, ISBN-13: 978 0 85709 498]. IES (Institute for Environment and Sustainability). European Commission, Joint Research Centre. Characterisation factors of the ILCD Recommended Life Cycle Impact Assessment methods. Database and Supporting information1st ed. Publications Office of the European UnionLuxemburg: Publications Office of the European Union; 2012. IES (Institute for Environment and Sustainability). European Commission, Joint Research Centre. International Reference Life Cycle Data System (ILCD) Handbook – General guide for Life Cycle Assessment – Detailed guidance. 1st ed. Luxemburg: Publications Office of the European Union; 2010. IDF (International Dairy Federation). A common carbon footprint approach for dairy – The IDF guide to standard lifecycle assessment methodology for the dairy sector. Bull Int Dairy Fed 2010;445:13. IPCCIn: Egglestone S, Buendia L, Miwa K, Ngara T, Tanabe K, editors. Guidelines for National Greenhouse Gas Inventory, Volume 4, Chapter 10. Prepared by the National Greenhouse Gas Inventories Programme. Hayama, Japan: Institute for Global Environmental Strategies; 2006. IPCCIn: Solomon S, Qin D, Manning M, Chen Z, Marquis M, Averyt KB, Tignor M, Miller HL, editors. Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge, United Kingdom: Cambridge University Press; 2007. ISO. Environmental Management – Life Cycle Assessment – Principles and Framework. EN ISO 14040Brussels, Belgium: EN ISO 14040; 2006a. ISO. Environmental Management – Life Cycle Assessment – Requirements and Guidelines. EN ISO 14044Brussels, Belgium: EN ISO 14044; 2006b. Kristensen G, Jensen JK. Emission factors for gas fired CHP units b25 MW. Emission mapping project reports for CHP plants less than 25 MW electrica; 2001 [www.neri.dk (last accessed 13 September 2013)]. Kristensen T, Mogensen L, Knudsen MT, Hermansen JE. Effect of production system and farming strategy on greenhouse gas emissions from commercial dairy farms in a life cycle approach. Livest Sci 2011;140:136–48. Li Zifu, Yin Fubin, Li Haoyuan, Wang Xiaoxi, Lian Jing. A novel test method for evaluating the methane gas permeability of biogas storage membrane. Renew Energy 2013;60: 572–7. Liebetrau J, Clemens J, Cuhls C, Hafermann C, Friehe J, Weiland P, et al. Methane emissions from biogas-producing facilities within the agricultural sector. Eng Life Sci 2010;10(2010):595–9. Loria ER, Sawyer JE. Extractable soil phosphorus and inorganic nitrogen following application of raw and anaerobically digested swine manure. Agron J 2005;97:879–85. Lyngsø Foged H, Flotats X, Bonmati Blasi A, Palatsi J, Magri A, Schelde KM. Inventory of manure processing activities in Europe. Technical Report No. I concerning “Manure Processing Activities in Europe” to the European Commission, Directorate-General Environment; 2011. [138 pp.]. Maranon E, Salter AM, Castrillon L, Heaven S, Fernandez-Nava Y. Reducing the environmental impact of methane emissions from dairy farms by anaerobic digestion of cattle waste. Waste Manag 2011;31:1745–51. Massé DI, Croteau F, Patni NK, Masse L. Methane emissions from dairy cow and swine manure slurries stored at 10° and 15 °C. Can Biosyst Eng 2003;45:61–6. Ministero dello Sviluppo Economico (MSE), Decreto 6 luglio 2012: Attuazione dell'art. 24 del decreto legislativo 3 marzo 2011, n. 28, recante incentivazione della produzione di energia elettrica da impianti a fonti rinnovabili diversi dai fotovoltaici. http:// www.sviluppoeconomico.gov.it/images/stories/normativa/DM_6_luglio_2012_sf.pdf (last access 13 September 2013.) Mircea M, Ciancarella L, Briganti G, Calori G, Cappelletti A, Cionni I, et al. Assessment of the AMS-MINNI system capabilities to simulate air quality over Italy for the calendar year 2005. Atmos Environ 2014;84:178–88. Möller K, Stinner W, Deuker A, Leithold G. Effects of different manuring systems with and without biogas digestion on nitrogen cycle and crop yield in mixed organic dairy farming systems. Nutr Cycl Agroecosyst 2008;82:209–32. Nguyen TTH, Doreau M, Corson MS, Eugène M, Delaby L, Chesneau G, et al. Effect of dairy production system, breed and co-product handling methods on environmental impacts at farm level. J Environ Manage 2013;120:127–37. Oenema O, Velthof GL, Verdoes N, Groot Koerkamp PWG, Bannink A, Monteny GJ, et al. Forfaitaire waarden voor gasvormige stikstofverliezen uit stallen en mestopslagen. AlterraWageningen UR Wageningen: Alterra; 2000. Opio C, Gerber P, Mottet A, Falcucci A, Tempio G, MacLeod M, et al. Greenhouse gas emissions from ruminant supply chains – A global life cycle assessment. Food and Agriculture Organization of the United Nations (FAO)Rome: Food and Agriculture Organization of the United Nations (FAO); 2013. PE International. Gabi 6 [Software]. Leinfelden-EchterdingenGermany: LeinfeldenEchterdingen; 2013. Perego A, Basile A, Bonfante A, De Mascellis R, Terribile F, Brenna S, et al. Nitrate leaching under maize cropping systems in Po Valley (Italy). Agric Ecosyst Environ 2012;147: 57–65. Posch M, Seppälä J, Hettelingh JP, Johansson M, Margni M, Jolliet O. The role of atmospheric dispersion models and ecosystem sensitivity in the determination of characterisation factors for acidifying and eutrophying emissions in LCIA. Int J Life Cycle Assess 2008;13:477–86.

208

F. Battini et al. / Science of the Total Environment 481 (2014) 196–208

Rabl A, Spadaro JV. The RiskPoll software, version is 1.051 (dated August 2004); 2004 [www.arirabl.com]. Rossi P, Gastaldo A. Consumi energetici in allevamenti bovini da latte. L'informatore agrario 2012;3:45–7. Rossier D. Ecobilan. Adaptation de la méthode écobilan pour la gestion environnementale de l'exploitation agricole. Service Romand de Vulgarisation AgricoleLausanne, Switzerland: Service Romand de Vulgarisation Agricole; 1998. Sanz-Cobena A, Sánchez-Martín L, García-Torres L, Vallejo A. Gaseous emissions of N2O and NO and NO3 − leaching from urea applied with urease and nitrification inhibitors to a maize (Zea mays) crop. Agric. Ecosyst. Environ. 2012;149:64–73. Schievano A, D'Imporzano G, Salati S, Adani F. On-field study of anaerobic full-scale plants (Part I): an on-field methodology to determine mass, carbon and nutrients balance. Bioresour Technol 2011;102:7737–44. Schröder J, Uenk D. Cattle slurry digestion does not improve the long term nitrogen use efficiency of farms. 12th RAMIRAN International Conference, “Technology for Recycling of Manure and Organic Residues in a Whole Farm Perspective”, RAMIRAN 2006, Aarhus, Denmark. Vol. II; 2006. p. 9–11. Seppälä J, Posch M, Johansson M, Hettelingh JP. Country-dependent Characterisation Factors for Acidification and Terrestrial Eutrophication Based on Accumulated Exceedance as an Impact Category Indicator. Int J Life Cycle Assess 2006;11(6):403–16. Smith KA, Cumby T, Lapworth J, Misselbrook TH, Nigro E, Williams AG. Natural crusting of slurry storage as an abatement measure for ammonia emissions on dairy farms. Biosyst Eng 2007;97:464–71. Sneath RW, Beline F, Hilhorst MA, Peu P. Monitoring GHG from manure stores on organic and conventional dairy farms. Agric Ecosyst Environ 2006;112:122–8.

Sonesson U, Cederberg C, Berglund M. Greenhouse gas emissions in milk production. Decision support for climate certification. Report 2009; 2009. p. 3. Stehfest E, Bouwman L. N2O and NO emission from agricultural fields and soils under natural vegetation: summarizing available measurement data and modelling of global annual emissions. Nutr Cycl Agroecosyst 2006;74:1314–85. Tukker A, Huppes G, Guinée J, Heijungs R, de Koning A, van Oers L, et al. Environmental impacts of products (EIPRO). Analysis of the Life Cycle Environmental Impacts related to the final consumption of the EU-25. Technical Report EUR 22284 EN. European Commission, Joint Research Centre, Institute for Prospective Technological Studies. Seville, Spain; 2006. Van der Werf HMG, Kanyarushoki C, Corson MS. An operational method for the evaluation of resource use and environmental impacts of dairy farms by life cycle assessment. J Environ Manage 2009;90:3643–52. Weiske A, Vabitsch A, Olesen JE, Schelde K, Michel J, Friedrich R, et al. Mitigation of greenhouse gas emissions in European conventional and organic dairy farming. Agric Ecosyst Environ 2006;112:221–32. Winther M, Nielsen OK. Fuel use and emissions from non-road machinery in Denmark from 1985–2004 and projections from 2005–2030. National Environmental Research InstituteDenmark: National Environmental Research Institute; 2006. Yan MJ, Humphreys J, Holden NM. An evaluation of life cycle assessment of European milk production. J Environ Manage 2011;92:372–9. Zappavigna P. Consumi di energia in stalla, quanto si può risparmiare. L'informatore agrario 2010;48:40–3. Zhang G, Strøm JS, Li B, Rom HB, Morsing S, Dahl P, et al. Emission of ammonia and other contaminant gases from naturally ventilated dairy cattle buildings. Biosyst Eng 2005;92:355–64.