Contents lists available at ScienceDirect

Livestock Science

Livestock Science 161 (2014) 176–184

1871-14http://d

n CorrE-m

journal homepage: www.elsevier.com/locate/livsci

Carbon footprint of milk produced at Italian buffalo farms

G. Pirlo a,n, G. Terzano b, C. Pacelli c, F. Abeni a, S. Carè a

a CRA-FLC, Consiglio per la ricerca e sperimentazione in agricoltura, Centro di ricerca per le produzioni foraggere e lattiero-casearie,Via Porcellasco 7, 26100 Cremona, Italyb CRA-PCM, Consiglio per la ricerca e sperimentazione in agricoltura, Centro di ricerca per la produzione delle carni e il miglioramentogenetico, Via Salaria 31, 00016 Monterotondo, Italyc Dipartimento di Scienze delle Produzioni Animali, Università della Basilicata, Campus of Macchia Romana, 85100 Potenza, Italy

a r t i c l e i n f o

Article history:Received 19 March 2013Received in revised form6 December 2013Accepted 8 December 2013

Keywords:Carbon footprintBuffalo milk

13/$ - see front matter & 2013 Elsevier B.V. Ax.doi.org/10.1016/j.livsci.2013.12.007

esponding author. Tel.: þ39 0372433029; faail address: giacomo.pirlo@entecra.it (G. Pirl

a b s t r a c t

The carbon footprint (CF) of milk produced in six Italian Mediterranean Buffalo farms wasestimated through a simplified Life Cycle Assessment. Functional unit was 1 kg of fat andprotein corrected milk (FPCM).The farms were characterized by high levels of inputs, aspurchased feeds, chemical fertilizers and fossil fuels. Average cultivated area was 53.2 ha;the forage system was based mainly on maize silage, immediately followed by Italianryegrass and/or whole cereal silage. Average herd size was 360 and the average FPCM perlactating buffalo was 3563 kg/year with an average milk fat and protein percentage of 8.24and 4.57 respectively. The CF assessment was from cradle to farm gate. The greenhousegases (GHG) that were taken into account were CH4 from enteric fermentations, CH4 frommanure in the stable and in the tank; N2O from nitrification and denitrification processesin the manure before application into the soil and N2O produced after organic andsynthetic fertilizer application; direct emissions of CO2 from the fossil fuels combustionwithin the farms and indirect emissions of CO2 deriving from production of electricity, off-farm feeds, synthetic fertilizers and other minor inputs. Carbon footprint of 1 kg of FPCMwas 3.75 kg CO2eq. Main sources of GHG are enteric CH4 (45%) and indirect CO2eq (25%).Besides enteric CH4, the farm activity that gives the highest contribution to milk CF ison-farm feed production, with 34% on total greenhouse gas emissions (TGE). Carbonfootprint with economic allocation (CFea) was estimated by considering the live-weight ofmales calves and culled cows: its value was 3.60 kg CO2eq. If the economic value of theincrease of the herd size is considered in the assessment, CFea decreases to 3.45 or 3.27 kgCO2eq with an increase of 10% or 20% of the number of mature buffalos.

& 2013 Elsevier B.V. All rights reserved.

1. Introduction

Countries that signed the Kyoto Protocol (1997) estab-lished to reduce the atmospheric emissions of greenhousegases (GHG) estimated in 1990. In the International Panel ofClimate Changes (IPCC) framework, GHG attributed to theagricultural sector are CH4 and N2O. Methane is producedmainly with enteric fermentation and decomposition of

ll rights reserved.

x: þ39 0372435056.o).

manure organic matter; N2O derives from the N content ofmanure and from N of organic and chemical fertilizers oncethey are applied to the soil. Technological progress and theadoption of best practices made possible to reduce GHGemissions from livestock production systems: in Italy, from1990 to 2009, emissions of enteric CH4, manure CH4, manureN2O, and soil N2O emissions decreased by 11.5%, 16.6%, 4.03%,and 20.6% respectively (Cóndor, 2011). However, CO2 has alsoa great interest for agriculture; in fact, this gas is emittedfrom the fossil fuel combustion of engines used for farmingoperations or from industrial activities supplying goods suchas chemical fertilizers or equipment.

G. Pirlo et al. / Livestock Science 161 (2014) 176–184 177

Life Cycle Assessment (LCA) is a widely acceptedmethodology for estimating the environmental effects thata unit of product causes during the entire life cycle. Carbonfootprint (CF) is an LCA referring exclusively to GlobalWarming Potential (GWP).

In Italy, there is an important population of dairybuffalos, whose milk is used for the production of “Moz-zarella di bufala campana – DOP”. The population of ItalianMediterranean Buffalos (IMB) has increased considerablyin the last decades; between 1985 and 2010 the number ofbuffalo cows increased from 45,000 to 245,000 (AIA,2010). However, in spite of the relevant economic impor-tance of these animals in some southern regions of Italyand of the worldwide spread of buffalos, so far we do notknow any specific study about CF of buffalo milk.

On the contrary, there are several studies concerning CFof cattle milk, as reviewed by Pirlo (2012). In order toharmonize the studies and make them comparable, theInternational Dairy Federation (FIL-IDF, 2010) proposed astandardization of the methodology for estimating the CFof 1 kg of fat and protein corrected milk (FPCM). FAO(2010) calculated that the average emission of GHG at thefarm-gate is about 2.4 kg of carbon dioxide equivalent(CO2eq) per kg of FPCM, with wide variations dependingon the geographic area: from 1 kg of CO2eq in NorthAmerica to 7.5 kg of CO2eq in South Africa. The productionsystem has also a great influence on milk CF: Capper et al.(2009) demonstrated that technological progress, occurredbetween 1944 and 2007, determined a reduction from 3.66to 1.35 kg CO2eq per kg of milk. The main driver of thispattern is the output of milk per cow per year (Gerberet al., 2011), that makes it possible to produce the sameamount of milk with a reduced number of cows.

In IMB population, the introduction of rational feedingtechnologies and improved management has also deter-mined a strong increase of average milk production; from1990 to 2012 the average milk production of buffalo cowsof the IMB Herdbook increased from 1893 to 2218 kg peryear, while milk fat and protein contents remained steady:about 8.2% and 4.5% respectively (ANASB, 2013). Rearingsystem is very similar to that of highly producing dairycattle and the diet is commonly based on maize silage,concentrates, alfalfa hay and a minor portion of by-products(Borghese and Mazzi, 2005).

Milk production accounts for the greatest part of theenvironmental burden of milk end products, such as drink-ing milk, cream, butter, or cheese (Fantin et al., 2012; Flysjö,2011; González-García et al., 2013); consequently, a reduc-tion of the GHG emissions in the agricultural phase can havepositive effects on the environmental impact of the whole

Table 1Forage system.

Farm 1

Total cultivated area (ha) 38Maizeþ Italian ryegrass or whole cereal silage (ha) 36Synthetic N fertilizers (N kg/ha) 349P fertilizers (P2O5 kg/ha) 70.8

In farm 3 the animals have access to pasture.

production chain. However, similar considerations can bemade for buffalo milk too.

Aims of this study were: (i) to estimate the CF of buffalomilk at farm gate and (ii) to point out the main drivers thatinfluence it.

2. Material and methods

2.1. Simplification

A simplified LCA was used for evaluating the CF of onekilogram of milk produced in six buffalo farms. Mainsimplifications were: (i) dry matter intake was not mea-sured, but was estimated from buffalos0 nutrient require-ments; (ii) resource consumptions and GHG emissionscaused by on-farm production were not estimated foreach crop, but as a whole; and (iii) electric consumptionswere estimated as a whole, without considering everyspecific activity, such as milking, cooling, or lighting.

2.2. Data collection

The milk CF was estimated in 6 buffalo farms located inCampania region. The farms were chosen at random from thepopulation of farms assisted by the technical service in thatregion. The data, referring to 2010, were collected through aquestionnaire that was distributed among the farmers. Thequestionnaire concerned technical data about farm size, crops,herd size and herd composition, manure storage and proces-sing, agricultural operations, forage system, amounts and kindof seeds, chemical fertilizers, pesticides, plastics, detergents,medicines, fuels, electricity, equipment, agricultural machines,amounts of off-farm feeds, production of on-farm crops anddiet fed to the different animal categories. Farmers also filledin the questionnaire with milk composition and output ofmilk andmeat. The questionnaire also required economic dataconcerning the yearly revenue for milk and for culled orreplacement animals.

The farms adopted a very intensive forage system(Table 1), permitted by the great availability of irrigationwater. Maize silage was the main crop, immediatelyfollowed, after the harvesting, by Italian ryegrass and/orwhole cereal silages. Table 1 reports the levels of N and Pfertilizers per hectare. A small portion of cultivated area wasexploited for alfalfa hay production. Only in one farmbuffalos had access to pasture for a short period of the year.

Herd size and herd composition are reported in Table 2.Manures were handled and stored for at least 3 monthsbefore spreading.

Farm 2 Farm 3 Farm 4 Farm 5 Farm 6 Mean

24 45 157 20 35 53.221 39 3 18 34 25.2201 310 4 307 277 24180.3 78.8 60.5 78.2 79.0 74.6

Table 2Herd size and composition.

Farm1 Farm 2 Farm 3 Farm 4 Farm 5 Farm 6 Mean

buffalos (n) 465 365 254 460 220 401 360Mature buffalos (n) 200 175 141 276 105 262 193Young buffalos (n) 265 190 113 184 115 139 167

Table 3Milk and live-weight output.

Farm 1 Farm 2 Farm 3 Farm 4 Farm 5 Farm 6 Mean

Total milk output (kg/year) 413,545 210,324 190,550 375,435 290,500 372,500 308,809Milk fat (%) 8.30 8.50 8.00 7.63 8.50 8.50 8.23Milk protein (%) 4.45 4.90 4.60 4.31 4.60 4.50 4.56Average daily milk production (kg/buffalo) 10.21 4.97 7.76 6.59 11.95 9.26 8.45Total FPCM output (kg/year) 644,542 326,920 292,841 548,768 462,248 588,657 477,329Average FPCM per lactation (kg/lactating buffalo/year) 4297 2179 3218 2601 5136 3951 3564Total live-weight output (kg/year) 29,500 17,000 5900 29,500 12,450 31.500 20,975

G. Pirlo et al. / Livestock Science 161 (2014) 176–184178

Data concerning milk production and meat output arereported in Table 3. Milk output was expressed as FPCMwith a fat and protein percentages of 4 and 3.3 respec-tively. The following equation reported, as by Gerber et al.(2011), was used for conversion of raw milk to FPCM:

FPCM ðkgÞ ¼M½0:337þ0:116FCþ0:06PC�where M is the mass of raw milk (kg), FC is the percentageof milk fat, and PC is the percentage of milk protein. Meatoutput was calculated summing the live-weight (LW) ofcalves, heifers and culled buffaloes sold for replacement,fattening or slaughtering. Live-weight of calves and culledbuffalo cows were considered as 45 and 600 kg respec-tively. No replacement heifers were sold in the farmsduring the time period considered in the study. Almostall income from meat production was due to culledbuffalos, the price of which varied between 0.5 and1.37 €/kg LW. Buffalo milk production varied considerablyfrom spring – summer to autumn – winter period. Onaverage, spring – summer milk production was approxi-mately 25% of the total yearly amount. Because milk priceis conditioned by mozzarella cheese consumption, thereare seasonal milk prices, with higher values during springand summer and lower values during autumn and winter.The income frommilk production was calculated summingthe income from milk produced in the former period withthe income from the latter one. Synthetically, milk pricevaried between 1.37 and 1.54 €/kg and between 1.15 and1.26 €/kg in spring – summer and in autumn – winterperiod respectively.

Table 4 shows the composition of feed rations for thedifferent animal categories. The feeding system was basedon maize, Italian ryegrass and whole cereal silage, entirelyproduced on-farm; grass or alfalfa hay, coming from thesurrounding agricultural areas, contributed for 20% to 30%of the total dry matter consumption. The concentrates,whose main component was maize and barley from northItaly and soya-bean meal from South America, contributedfor 18% to 23% of the total dietary dry matter. In somefarms, straw was fed to pregnant heifers or dry buffaloes.

Table 5 shows the amounts of diesel, electricity, che-mical fertilizers and off-farm feeds which represent themain inputs of the six farms.

2.3. The system boundaries

The system boundaries (Fig. 1) comprised the GHGemissions associated to the production of buffalo milk upto the farm gate. Consequently, they did not consider GHGemissions during milk transportation to dairies, processingand retail. The system included: (i) direct emissions of CH4

and N2O associated with enteric fermentations and withmanure management, storage and application; (ii) directGHG emissions (as CO2eq) from fuel combustion for theworking of engines used in agricultural operations, manurehandling and milking; (iii) indirect emissions of GHG (asCO2eq), derived from the production of chemical fertilizers,seeds, pesticides, plastic, off-farm feeds production, transportand processing, fossil fuel extraction, production and main-tenance of agricultural equipment and machines. Emissionsderiving from the construction of farm facilities (e.g. barns,roads, and irrigation channels) were not considered. In thefarm sample, land use was under constant managementpractice, so that the emissions from land use change wereexcluded from the analysis. Carbon dioxide emission fromanimal respiration and from enteric fermentation was nottaken into account because it is not considered a net sourceunder the Kyoto Protocol. Carbon dioxide fixed by plantsthrough photosynthesis was excluded because equivalent toCO2 produced through animal respiration and enteric fer-mentation (IPCC, 2006).

The functional unit (FU) used to report GHG emissionwas defined as 1 kg of FPCM.

2.4. Allocation

Milk production gives more than one functional out-put; besides milk, there is meat from culled cows andsurplus calves. All these outputs were expressed as kilo-gram of LW. A number of options exist to apportion that

Table 4Contribution of feeds to diet composition and crude protein concentration (% of dry matter).

Farm 1 Farm 2 Farm 3 Farm 4 Farm 5 Farm 6 Mean

Lactating buffalosMaize silage 45 35 36 14 45 38 35.5Concentrate 22 20 21 5 22 22 18.7aGrass or legume hay 10 30 35 50 30 30 30.8Italian ryegrass or whole cereal or hay 23 15 8 31 3 10 15.0Crude protein concentration 16.5 15.0 16.5 14.0 17.0 15.5 15.7

Heifer less than 2 yearsMaize silage 45 40 38 12 42 39 36Concentrate 26 23 24 15 21 25 22.3aGrass or legume hay 8 19 30 34 20 18 21.5Italian ryegrass or whole cereal or hay 21 18 8 39 17 18 20.2Crude protein concentration 15.2 15.0 14.5 13.8 15.0 15.1 14.7

Pregnant heifersMaize silage 36 37 30 10 30 31 29Concentrate 19 18 25 32 22 20 22.7aGrass or legume hay 24 19 30 33 39 32 29.5Italian ryegrass or whole cereal or hay 6 8 7 25 9 7 10.2Straw 15.8 18 8 – – 10 8.6Crude protein concentration 14.8 15.5 15.9 15.5 14.5 14.0 15.0

Dry buffalosMaize silage 29 25 32 18 24 22 25Concentrate 15 18 14 28 20 21 19.3aGrass or legume hay 30 25 21 26 50 26 29.7Italian ryegrass or whole cereal or hay 6 8 7 28 6 6 10.2Straw 20 24 26 – – 25 15.8Crude protein concentration 12.3 13.6 13.0 15.2 14.0 13.5 13.6

a Pasture in farm 3.

Table 5Main inputs.

Farm 1 Farm 2 Farm 3 Farm 4 Farm5 Farm 6

Diesel 1000 L/year 28.2 18.0 18.2 44.3 57.0 15.6Electricity MWh/year 12.0 7.6 8.6 7.5 34.1 34.1Chemical fertilizers tonne/year 49.4 25.5 52.1 4.9 37.9 57.3Off-farm feeds tonne/year 83.3 205.0 82.1 132.0 87.3 140.2

Fig. 1. System boundaries.

G. Pirlo et al. / Livestock Science 161 (2014) 176–184 179

part (FIL-IDF, 2010). For the present analysis, the contribu-tion of co-products was estimated on the basis of theireconomic value (economic allocation), by using the afore-mentioned mass and economical values (see Section 2.2).

Some farms in the sample also produced excess manurewith respect to crop requirements. This mass was given tosome neighboring farms. Because no economic value wasattributed to the surplus manure, it was not accounted forin the economic allocation. However, the effect on milk CFwas estimated on the basis of manure N content, con-sidering the reduction of CO2 emissions deriving from theproduction, packaging and transport of a correspondingamount of synthetic fertilizers (see Section 2.6).

2.5. Impact category and characterization

The GHG emissions were expressed as GWP in a 100-year time horizon, defined as CO2eq: 1 kg CO2¼1 kgCO2eq; 1 kg CH4¼25 kg CO2eq; and 1 kg N2O¼298 kgCO2eq (Forster et al., 2007). The total GHG emissions(TGE), expressed as kg CO2eq, were determined as

TGE¼ ΣGWPi�mi

where mi is the mass in kg of the emitted gases.

Table 6Emission factors used in carbon footprint assessment.

Emission factor Unit References

CH4 enteric fermentationsMature buffalos aGEx(Ym/100)365/55.65 kg CH4/head/year IPCC (2006)Young buffalos 3 months-1 year 21.16 kg CH4/head/year Cóndor et al. (2008b)Young buffalos 1–3 years 73.42 kg CH4/head/year Cóndor et al. (2008b)

CH4 Manure managementGrazing 0.98 g CH4/kg DM of dung Wheeler (2012)Confined mature buffalos – liquid system 4.49 kg CH4/head/year Cóndor et al. (2008a)Confined mature buffalos – solid system 10.26 kg CH4/head/year Cóndor et al. (2008a)Other buffalos – liquid system 3.13 kg CH4/head/year Cóndor et al. (2008a)Other buffalos – solid system 3.16 kg CH4/head/year Cóndor et al. (2008a)

N2O direct emissionsSolid manure storage 0.02�N excreted kg N2O/kg N Cóndor et al. (2008a)Liquid manure storage 0.001N excreted kg N2O/kg N Cóndor et al. (2008a)Synthetic N and organic fertilizers and N-fixing crops 0.0125 kg N–N2O/kg N Cóndor et al. (2008a)N2O indirect emissionsAtmospheric N redeposition 0.01 kg N2O/kg N Cóndor et al. (2008a)Leaching-runoff 0.025 kg N2O/kg N Cóndor et al. (2008a)

CO2 directDiesel combustion 3.13 kg CO2/kg APAT (2003)Liquid propane gas 2.872 kg CO2/kg APAT (2003)

CO2 indirectDiesel extraction, refining and transportation 1.0062 kg CO2/kg O0Brien et al. (2010)Electricity 0.47 kg CO2/kWh Cóndor (2011)Production and transport of urea 1.29 kg CO2/kg Ceschia et al. (2010)Production and transport of ammonium nitrate 1.11 Ceschia et al. (2010)Production and transport of phosphoric fertilizers 0.42 Ceschia et al. (2010)Production and transport of pesticides, seeds, plastic, machines Rotz and Chianese

(2009)Production and transport of concentrates and straw Ecoinvent (2009)CO2 surplus manure 0.95 kg CO2/kg N Ceschia et al. (2010);

CRPA (2001)

a GE: gross energy (MJ); Ym: percent of GE converted to CH4; 55.65: energy content of 1 kg of CH4.

G. Pirlo et al. / Livestock Science 161 (2014) 176–184180

2.6. Methodology

Table 6 shows the list of the emission factors (EF) usedfor calculating GHG emissions.

Methane emissions arising from enteric fermentationsof lactating and dry buffalos were estimated through aTier 2 methodology (IPCC, 2006). The EF was calculated asfollowing:

EF¼GEðYm=100Þ365=55:65where EF is the amount (kg) of CH4 emitted by a head peryear, GE is the gross energy intake (MJ per head per day)calculated on the basis of net energy requirements formaintenance, activity, growth, lactation, and pregnancy ofdairy and buffalo cows (NRC, 2001). The factor 55.65 MJ isthe energy content of 1 kg of CH4. The percentage of GEconverted to CH4 (Ym) was assumed to be equal to 6 foradult buffaloes. Parameters used to estimate the EF ofenteric CH4 of other buffalo categories were 21.16 kg CH4/head/year for calves between 3 months and 1 year and73.42 kg CH4/head/year for heifers between 1 and 3 years(Cóndor et al., 2008b).

Methane emissions from manure management arisingduring storage and following application on land was esti-mated using the IPCC methodology (IPCC, 2006) adapted toItalian condition by Cóndor et al. (2008a). Because of lack ofcountry-specific parameters, the EF of 0.98 g of CH4/kg of DMof dung was used in case of grazing (Wheeler, 2012).

According to IPCC (2006), emission of N2O was calculatedfrom N excretion and specific EFs, depending upon manuremanagement. Nitrogen excreted by lactating buffalos wasestimated by calculating N intake and subtracting N retained.Dry matter intake was the same as that calculated forestimating enteric CH4; this value was successively multi-plied by the diet N content. The amount of N retained wascalculated by considering milk production and a referenceLW gain of 0.055 kg/d (Cóndor et al., 2008b). Nitrogenexcreted from other buffalo categories was estimated accord-ing to IPCC (2006) by adopting the Italian country-specificvalue of 16.2 kg N/head per year (Cóndor et al., 2008a).Nitrous oxide emissions associated with manure storage andmanagement were calculated multiplying total N excretionby Italian country-specific EF of 0.001 kg of N–N2O/kg Nexcreted (Cóndor et al., 2008a). Direct soil N2O emissionswere computed on the basis of the N applied to the soil withorganic and synthetic fertilizers, crop residues and N-fixingcrops, adjusted for the N lost through volatilization as NH3 orNOx. Intergovernmental Panel Climate Change (2006) meth-odology was also used for calculating indirect N2O emissions,following nitrate leaching and run-off, and followingre-deposition of volatized gases to soils and waters. Emissionfactor of N2O derived from direct N soil deposition was0.0125 kg N–N2O/kg N; for atmospheric indirect depositionand leaching-runoff, EF were 0.01 kg N2O/kg N and 0.025N–N2O/kg N respectively (country-specific EF obtained fromCóndor et al. (2008a)).

G. Pirlo et al. / Livestock Science 161 (2014) 176–184 181

Fuel used for agricultural operations and animal feedingwas estimated by summing the amount of purchased fuelwith that used by contractors, which was calculated resort-ing to the Handbook of fuel consumption in agriculture(ENAMA, 2005). A value of 0.85 kg per liter was consideredfor diesel density. Emission factors for the combustion of1 kg of diesel and 1 kg of liquid propane gas (LPG) are 3.13and 2.872 CO2eq (APAT, 2003). For extraction, refining, andtransport of diesel an EF of 1.0062 kg of CO2eq per L wasused (O0Brien et al., 2010); the EF for energetic consumptionconsiders not only CO2 but also minor amounts of otherGHGs produced along the chain from extraction throughoututilization.

Emission factor associated with the production of 1 kWhof electricity, used for milking, milk cooling, barn lighting,ventilation, and other farm activities, is 0.47 kg CO2eq(Cóndor, 2011). Emission factors corresponding to manufac-turing, packaging, transport and storage of mineral fertilizerswere those reported by Ceschia et al. (2010).

Greenhouse gas emissions from the production ofpesticides, seeds, plastic and machine were estimatedfollowing the indications of Rotz and Chianese (2009).Weight of plastic film normally used for sealing silagetrenches was considered 0.11 kg/m2. Weight of net usedfor silage or hay round balls was 33 kg every 200 m oflength. Greenhouse gas emissions from the production ofpesticides were estimated using the average EF of 22 kg ofCO2eq per kg of pesticide. Seed use factors were derivedfrom typical seeding rates and yields of each crop. Anaverage EF of 0.3 kg of CO2eq/kg of seed was used. Forbottle plastic the EF was 2 kg of CO2eq/kg of plastic.Greenhouse gas emissions produced during machine man-ufacture were calculated on the basis of total mass of farmmachines, using the EF of 3.54 kg of CO2eq/kg of machine

Table 7Carbon dioxide equivalents emissions (kg CO2eq/year).

Farm 1 Farm 2 Farm 3

CH4 enteric 825,946 633,939 545,06CH4 manure managment 74,336 44,077 54,07N2O manure managment 178,527 122,860 3N2O synthitic fertilizers 94,344 77,408 78,62N2O manure soil 170,220 134,320 105,63N2O crop. res and N-fix. crops 9888 10,334 11,21N2O indirect emissions 160,712 220,814 125,26CO2 fuels direct emissions 73,992 48,076 48,50CO2 fuel indirect and electricity 41,874 32,247 28,76CO2 synth. fertilizer production 110,008 47,190 150,88CO2 seed, pesticides, plastic, equip. prod. 21,772 15,428 18,28CO2 off-farm feeds 166,955 340,921 161,87Total gross emissions 1,928,577 1,75,695 1,358,24Surplus manure �77,143 �17,157Total net emissions 1,851,434 1,698,538 1,358,24

Table 8Carbon footprint of 1 kg of FPCM (kg CO2eq).

Farm 1 Farm 2

Carbon footprint 2.87 5.20Carbon footprint with economic allocation 2.74 4.91

mass (Rotz and Chianese, 2009). Amount of CO2eq asso-ciated with the production, processing and transport ofconcentrate feed and straw was calculated utilizing massallocation criteria for subdividing GHG gas emissionsamong main and co-products (Ecoinvent, 2009).

Inputs, outputs, mass emissions and correspondingkilograms of CO2eq were calculated through an electronicsheet (Pirlo and Carè, 2013).

3. Results and discussion

3.1. Carbon footprint

Table 7 shows the productions of GHG, expressed asCO2eq, from different source and activities in the sixbuffalo farms. The results concerning GHG emissionsassociated with the production of 1 kg of FPCM have beenreported on Table 8; the values referred to milk CF withoutand with economic allocation (CFea); allocation factor was4.06% of TGE.

3.2. Contribution of different sources and activities to milkcarbon footprint

The contribution of each source to CF of cattle andbuffalo FPCM are shown in Fig. 2. The two main sourceswere enteric CH4 and indirect emissions of CO2eq, whichrepresented about 45% and 25% of CF respectively.

Enteric CH4 is a by-product of the fibre-hydrolysis andfermentative processes taking place in the four stomachs ofruminants, allowing these animals to obtain fatty acids andprotein from roughages. Some authors used different Ym asfunction of feeding regimen to consider the effect of dietcomposition on enteric CH4 production; in this respect,

Farm 4 Farm 5 Farm 6 Mean

0 778,991 860,190 890,877 755,8332 100,153 3114 6.875 47,1049.079 195,129 35,580 19,749 98,4879 18,225 53,160 75,969 66,2896 0.000 143.709 123.745 112,9381 3806 10,074 6016 85552 160.729 160,458 112,308 156,7149 120,826 155,390 39,100 80,9825 92,773 91,525 50,128 56,2195 5709 62,728 130,353 84,5993 11,952 12,722 19,047 16,5349 270,857 175,367 360,482 246,0772 1,742,839 1,739,182 1,813,581 1,716,3490 0 �86,959 0 �30,2102 1,742,839 1,652,223 1,813,561 1,686,140

Farm 3 Farm 4 Farm 5 Farm 6 Mean

4.60 3.19 3.57 3.08 3.754.38 3.09 3.49 2.95 3.6

-10.00

0.00

10.00

20.00

30.00

40.00

50.00

Enteric

CH4Man

ure

CH4 Manure

N2O Direct

soil

N2O Indire

ct

soil N

2O Direct

CO2eq

Indire

ct

CO2eq

Surplus

manure

% o

n to

tal C

O2e

q

Fig. 2. Contribution of different sources of GHG to milk CF.

0

5

10

15

20

25

30

35

40

45

50

Enteric CH4 Manure mang.animal housing

and care

On-farm feeds Off-farm feeds

% o

n to

tal C

O2e

q

Fig. 3. Contribution of different activities to milk CF.

G. Pirlo et al. / Livestock Science 161 (2014) 176–184182

Nguyen et al. (2012) applied a reduction factor of 4.8% (gramCH4/kg of dry matter intake) per percentage unit of addedlipids to Ym and Yan et al. (2010) demonstrated that Ymlinearly decreases from 8.49% at maintenance feeding levelto 4.45% at a feeding level of 5.

Several studies concerning different feeding strategiesfor reducing enteric CH4 emissions have been thoroughlyreviewed by de Boer et al. (2011). These techniques can beadopted for IMB too, due to the great likeness with theintensive dairy cattle production system.

Capper et al. (2009) showed that, since 1944, theincrease in milk production per head in dairy cows wasaccompanied by a dilution of maintenance requirements.A similar observation can be made on the sample of IMBfarms. For this reason, the increase of milk production perbuffalo or the improvement of feed efficiency are undoubt-edly effective strategies for reducing enteric CH4 emissionper kilograms of FPCM, similarly to what has been pro-posed for dairy cattle (Wall et al., 2009). However, someauthors observed that the relationship between milkproductivity and CF could be influenced by the animalproduction level: Vellinga et al. (2011) reported that CFwas significantly decreased until a level of productivity of6500 kg of milk per year; over this level milk productivitydid not have much effect on CF. The increased rate of milkproduction of IMB (ANASB, 2013) is still very high, so thatfurther improvements are possible.

Manure is another source of CH4 (Fig. 2). Its contribu-tion to TGE was not very high (less than 3%), even thoughthe management systems met in the IMB considered farmswere favorable for manure CH4 emission. Since the littleimportance of manure CH4 in milk CF, there are fewfeasible strategies for reducing this source.

Manure N2O represents about 5.6% of TGE (Fig. 2). Twoare the factors determining emission of this gas frommanure: amount of N excreted and the manure managementsystem. Nitrogen losses per kilogram of milk can be reducedthrough an improvement of productivity (Powell et al., 2013)or of diet protein efficiency (Agle et al., 2010; Robinson,2010). In the considered IMB farms, there is not enoughroom for further reduction of N2O emission through themanagement system, because the liquid system creates an

aerobic condition that is already unfavorable for N2O pro-duction (Amon et al., 2006); however there is a margin tominimize the animal N excretion by reducing diet proteinconcentration or even by optimizing the rumen degradableand undegradable protein ratio.

Direct and indirect soil N2O emission contributes withover 14% of the TGE in the sample of IMB farms (Fig. 2).This gas is produced through nitrification and denitrifica-tion processes of N in the soil and is mainly determined byanimal and fertilizer N deposition. Consequently, the highcontribution of soil N2O to CF estimated in the sample ofbuffalo farms is attributable to high animal density andsynthetic N fertilizer rates.

Another key source of GHG is CO2 (Fig. 2). Direct CO2

emission derives mainly from diesel combustion for agri-cultural operations. Indirect emissions, instead, are attri-butable to inputs that produce GHG outside the farm; thisis the case of purchased concentrates, synthetic fertilizer(mainly N), seeds, pesticides, plastics, and machine man-ufacturing. This list also comprises CO2 and other GHG, asCH4 and N2O, emitted for producing electricity and for oilextracting and refining. In this study we found that directCO2 emission contributed with about 5.1% of TGE; a muchhigher value was estimated for indirect CO2 emission(about 25%). Such high value is often associated to inten-sive dairy production systems, where there is a greatconsumption of synthetic fertilizers and off-farm feeds;highly productive crops, typical of the IMB system, requirealso many agricultural operations and an elevated con-sumption of fossil fuel.

A significant amount of surplus of manure wasrecorded in this study. As this mass is utilized for produ-cing vegetables in neighboring farms, a considerableamount of synthetic fertilizers can be potentially saved.The utilization of surplus of manure corresponds to anaverage reduction of 1.74% of TGE.

Fig. 3 gives an overview about the contribution ofdifferent farm activities to milk CF. It shows that, besidesenteric CH4, the activity that contributes most to milk CF ison-farm feeds production, which comprises a long list ofdirect or indirect GHG sources: use and production ofchemical fertilizers, manure application, use and produc-tion of fossil fuels, and use and production of all the otheritems necessary for cropping, such as seeds, pesticides,

Table 9Simulation of buffalo heifers virtually exceeding culling rate and corre-sponding CFea of 1 kg of buffalo FPCM.

Percentage ofexceedingheifers

Exceedingheifers (n)

Allocationfactor

Economic allocationCF (kg CO2/kg FPCM)

10 19.2 11.6 3.3220 38.7 16.0 3.15

G. Pirlo et al. / Livestock Science 161 (2014) 176–184 183

plastic and equipment. This observation makes even moreevident that a strategy for reducing GHG emission frombuffalo milk production should consider an efficient use ofchemical fertilizers in combination with use of manure, asseen previously.

3.3. Meat production

Meat, as LW, from surplus male calves and heifers,culled cows and fattened bulls, is a co-product of milkproduction.

We calculated an economic allocation factor based onthe total sale of LW and on market price corresponding toeach animal category. Allocation reduced CF by 4.06%. TheGHG emission associated with 1 kg of buffalo LW produc-tion in a dairy system was 3.70. This value is very close tothose of Kristensen et al. (2011) for dairy cattle, whoreported 3.41 and 3.52 kg CO2eq for conventional ororganic farms respectively.

However, we have to observe that our approach did notconsider the yearly variation in herd size and herd composi-tion, which could have a great effect on CF. It is widelyrecognized that the culling rate is very high and longevity ispoor in intensive dairy cattle farms. This condition puts theintensive dairy farms substantially in a steady-state, becausethe herd size does not increase unless replacement heifersare purchased on the market. Instead, buffaloes are verylong-lived, if compared with highly producing dairy cattle.More precisely, IMB population shows an average culling ratelower than 15% (Zicarelli, 2010) allowing a fast increase ofherd size or, in alternative, the selling of a lot of replacementheifers. This characteristic of IMB population is witnessed byits considerable increase in the last decades (AIA, 2010).Although not recorded, we know that all the buffalo farmsthat we visited increased the number of their animalsthroughout 2010. For this reason we think that an economicallocation which considers only the animals that leave thefarm, does not estimate correctly CF of buffalo milk. We alsoneed to include heifers that contribute to increase the herdsize, by considering them virtually getting out of the systemboundaries. In practice, we should estimate them as exceed-ing heifers and calculate the corresponding revenue andestimate again CFea.

With this aim, we simulated an increase rangingbetween 10% and 20% of mature buffaloes. The corre-sponding number of animals, allocation factors and cor-rected economic allocated CFea are reported in Table 9.The contribution of meat production to CFea reaches 11.6%with an enhancing of 10% of herd size or it reaches 16.0%with an enhancing of 20%. Consequently, CFea of 1 kg ofFPCM is 3.45 and 3.27 kg of CO2eq if the quota of virtuallyexceeding heifers is 10% or 20% respectively.

4. Conclusion

In this study we have examined the characteristics ofsix dairy buffalo farms. The production system is veryintensive, requiring a great amount of purchased feeds,chemical fertilizers and fossil fuels. The GHGs that con-tribute most to CF are consequently enteric CH4, typical ofruminants, and CO2 from indirect sources, similarly to

what can be commonly observed in intensive dairy cattlesystems. It has also been observed that the activity thatgives the greatest contribution to GHG gas emission, inaddition to enteric CH4, is off-farm feeds production, dueto the considerable amounts of fossil fuel and chemicalfertilizers that are required in an intensive productionsystem, such as that we have examined.

We adopted an economic criterion for allocating the meatassociated to milk production. However, analysis cannot belimited to animals that really leave the boundary systems, butit should consider also heifers that exceed the culling rate.

We have to remark the lack of specific EF for buffalosand of specific experiments for testing the efficacy ofmitigation measures. We think that this paper could beconsidered the initial step for a more detailed study of thisproduction system.

Conflict of interest

We confirm that in the article “Carbon footprint of milkproduction at Italian buffalo farms”: there are no conflict ofinterest associated with this publication; the work describedhas not been published previously and it is not underconsideration for public elsewhere; the manuscript has beenread and approved by all the named authors; and if accepted,it will not be published elsewhere without the writtenconsent of the copyright-holder.

Acknowledgment

This study has been funded through the Italian Ministryof Agricultural, Food, Forestry Policies (MiPAAF) “Sviluppodi modelli zootecnici ai fini della sostenibilità – SOS-ZOOT”.We also thank Dott. Paolo Masoni and coworkers for theirworthful advices.

References

Agle, M., Hristov, A.N., Zaman, S., Schneider, C., Ndegwa, P., Vaddella, V.K.,2010. The effects of ruminally degraded protein on rumen fermenta-tion and ammonia losses from manure in dairy cows. J. Dairy Sci. 93,1625–1637.

Amon, B., Kryvoruchko, V., Aman, T., Zechmeister-Boltenstern, S., 2006.Methane, nitrous oxide and ammonia emissions during storage andafter application of dairy cattle slurry and influence of slurry treat-ment. Agric. Ecosyst. Environ. 112, 153–162.

AIA, Associazione Italiana Allevatori, 2010. Quadro economico e relazioneattività 2010.

ANASB, Associazione Nazionale Allevatori Specie Bufalina, 2013. Availablefrom: ⟨http://www.anasb.it/dati-nazionali.php⟩. (accessed 10.05.13).

APAT, Agenzia per la protezione dell0ambiente e per i servizi tecnici, 2003.Analisi dei fattori di emissione di CO2 dal settore dei trasporti.Rapporti 28/2003.

G. Pirlo et al. / Livestock Science 161 (2014) 176–184184

Borghese, A., Mazzi, M., 2005. Buffalo population and strategies in theworld. In Buffalo production and research. In: Borghese, A. (Ed.), ,Food and Agriculture Organization of the United Nations, Rome, Italy,pp. 1–15.

Capper, J.L., Cady, R.A., Bauman, D.E., 2009. The environmental impact ofdairy production: 1944 compared with 2007. J. Anim. Sci. 87,2160–2167.

Ceschia, E., Béziat, P., Dejoux, J.F., Aubinet, M., Bernhofer, C.h., Bodson, B.,Buchmann, N., Carrara, A., Cellier, P., Di Tommasi, P., Elbers, J.A.,Eugster, W., Grünwald, T., Jacobs, C.M.J., Jans, W.W.P., Jones, M.,Kutsch, W., Lanigan, G., Magliulo, E., Marloie, O., Moors, E.J., Mour-eaux, C., Olioso, A., Osborne, B., Sanz, M.J., Saunders, M., Smith, P.,Soegaard, H., Wattenbach, M., 2010. Management effects on netecosystem carbon and GHG budgets at European crop sites. Agric.Ecosyst. Environ. 139, 363–383.

Cóndor, R.D., 2011. Agricoltura: emissioni nazionali in atmosfera dal 1990al 2009. Istituto superiore per la protezione e la ricerca ambientale,Rapporto ISPRA 140/2011. Roma, Italy.

Cóndor, R.D., Di Cristofaro, E., De Lauretis, R., 2008a. Agricoltura:inventario nazionale delle emissioni e disaggregazione provinciale.Istituto superiore per la prevenzione e la ricerca ambientale, ISPRARapporto tecnico 85/2008. Roma, Italy.

Cóndor, R.D., Valli, L., De Rosa, G., Di Francia, A., De Lauretis, R., 2008b.Estimation of the methane emission factor for the Italian Mediterra-nean buffalo. Animal 2, 1247–1253.

CRPA, Centro Ricerche Produzioni Animali, 2001. Liquami zootecnici.Manuale per l0utilizzazione agronomica. Edizioni L0InformatoreAgrario, Bologna, Italy.

De Boer, I.J.M., Cederberg, C., Eady, S., Gollnow, S., Kristensen, T., Macleod,M., Meul, M., Nemecek, T., Phong, L.T., Thoma, G., van der Werf, H.M.G., Williams, A.G., Zonderland-Thomassen, M.A., 2011. Greenhousegas mitigation in animal production: towards an integrated life cyclesustainability assessment. Cur. Opin. Environ. Sustain. 3, 423–431.

Ecoinvent, 2009. The Life Cycle Inventory data. Swiss Centre for Life CycleInventories.

ENAMA, Ente Nazionale per la Meccanizzazione Agricola, 2005. Pron-tuario dei consumi di carburante per l0impiego agevolato in agricol-tura. Roma 2005. Available from: ⟨http://www.enama.it/php/pageflip.php?pdf=enama_int_prontuario.pdf&dir=/it/pdf/monografie⟩.

Fantin, V., Buttol, P., Pergreffi, R., Masoni, P., 2012. Life cycle assessment ofItalian high quality milk production. A comparison with EDP study. J.Clean. Prod. 28, 150–159.

FAO, Food and Agriculture Organization of the United Nations, 2010.Greenhouse gas emissions from the dairy sector. A life cycle assess-ment. FAO, Rome, Italy.

FIL-IDF, International Dairy Federation, 2010. A carbon footprint approachfor dairy. The IDF guide to standard lifecycle assessment methodol-ogy for the dairy sector. Bull. Int. Dairy Fed. 445, 1–40.

Flysjö, A., 2011. Potential for improving the carbon footprint of butter andblend products. J. Dairy Sci. 94, 5833–5841.

Forster, P., Ramaswamy, V., Artaxo, P., Berntsen, T., Betts, R., Fahey, D.W.,Haywood, J., Lean, J., Lowe, D.C., Myhre, G., Nganga, J., Prinn, R., Raga,G., Schulz, M., Van Dorland, R., 2007. Changes in atmosphericconstituents and in radiative forcing. In Climate Change 2007: thephysical science basis. Contribution of the Working Group I to theFourth Assessment Report of the Interngovernmental Panel onClimate Change. In: Salomon, S., Qin, D., Manning, M., Chen, Z.,Marquis, M., Averyt, K.B., Tignor, M., Miller, H.L. (Eds.), , Cambridge

University Press, Cambridge, UK and New York, NY, USA ⟨http://www.ipcc.ch/pdf/assessment-report/ar4/wg1/ar4-wg1-chapter2.pdf⟩.

Gerber, P., Vellinga, T., Opio, C., Steinfeld, H., 2011. Productivity gains andgreenhouse gas emissions intensity in dairy systems. Livest. Sci. 139,100–108.

González-García, S., Castanheira, É.G., Dias, A.C., Arroja, L., 2013. Environ-mental performance of a Portoguese mature cheese-making dairymill. J. Clean. Prod. 41, 65–73.

IPCC, 2006. 2006 IPCC Guidelines for National Greenhouse Gas Inven-tories: volume 4: Agriculture, Forestry and other Land Use. Inter-governmental Panel on Climate Change. Available from: ⟨www.ipcc-nggip.iges.or.jp/public/2006gl/vol4.html⟩ (visited in February 2012).

Kristensen, T., Mogensen, L., Knudsen, T.M., Hermansen, J.E., 2011. Effectof production system and farming strategy on greenhouse gasemissions from commercial dairy farms in a life cycle approach.Livest. Sci. 140, 136–148.

Kyoto Protocol, 1997. Kyoto Protocol to the United States FrameworkConvention on Climate Changes. Available from: ⟨http://unfccc.int/resources/docs/convkp/kpeng.html⟩.

NRC, National Research Council, 2001. Nutrient Requirements of DairyCattle, Seventh revised edition, 2001 National Academic Press,Washington, D.C.

Nguyen, T.T.H., van der Werf, H.M.G., Eugène, M., Veysset, P., Devun,J., Chesneau, G., Doreau, M., 2012. Effects of type of ration andallocation methods on the environmental impacts of beef-production systems. Livest. Sci. 145, 239–251.

O’Brien, D., Shalloo, L., Grainger, C., Buckley, F., Horan, B., Wallace, M.,2010. The influence of strain of Holstein-Friesian cow and feedingsystem on greenhouse gas emissions from pastoral dairy farms. J.Dairy Sci. 93, 3390–3402.

Pirlo, G., 2012. Cradle-to-farm gate analysis of milk carbon footprint: adescriptive review. Ital. J. Anim. Sci. 11 (e20), 109–118.

Pirlo, G., Carè, S., 2013. A simplified tool for estimating carbon footprintfor dairy cattle milk. Ital. J. Anim. Sci. 12 (281), 497–506.

Powell, J.M., MacLeod, M., Vellinga, T.V., Opio, C., Falcucci, A., Tempio, G.,Steinfeld, H., Gerber, P., 2013. Feed-milk-manure nitrogen relation-ships in global dairy production systems. Livest. Sci. 152, 261–272.

Robinson, P.H., 2010. Impacts of manipulating ration metabolizable lysineand methionine levels on the performances of lactating dairy cows: asystematic review of the literature. Livest. Sci. 127, 115–126.

Rotz, C.A., Chianese, D.S., 2009. The dairy greenhouse gas model.Reference manual Version 1.2. Pasture System and Watershed man-agement Research unit. Agric. Res. Serv. United States Dep. Agric..

Vellinga, Th.V., de Haan, M.H.A., Schils, R.L.M., Evers, A., van den Pol-vanDasselaar, A., 2011. Implementation of GHG mitigation on intensivedairy farms: farmers0 preferences and variation in cost effectiveness.Livest. Sci. 137, 185–195.

Wall, E., Simm, G., Moran, D., 2009. Developing breeding schemes toassist mitigation of greenhouse gas emissions. Animal 4, 366–376.

Wheeler, D.M., 2012. OVERSEERs Technical manual. AGResearch Ltd.Yan, T., Mayne, C.S., Gordon, F.G., Porter, M.G., Agnew, R.E., Patterson, D.C.,

Ferris, C.P., Kilpatrick, D.J., 2010. Mitigation of enteric methaneemissions through improving efficiency of energy utilization andproductivity in lactating dairy cows. J. Dairy Sci. 93, 2630–2638.

Zicarelli, L., 2010. Enhancing reproductive performance in domestic dairywater buffalo (Bubalus bubalis). Soc. Reprod. Fertil. 67 (Suppl.),443–455.