Implications
Reducing greenhouse gas (GHG) emissions and increasing milk production is a major challenge facing the global dairy industry. However, to reduce GHG emissions per unit of milk, it is incorrect to consider only components of the dairy system relevant for policy reporting such as that used by the Intergovernmental Panel on Climate Change. Instead, approaches such as life cycle assessment, which consider on and off-farm GHG emissions should be used. This holistic approach ensures that emissions are reduced and not transferred to another location. Thus, reform of the present policy framework is needed to enable quantification of the impact of mitigation strategies on global GHG emissions.
Introduction
Globally, the dairy industry is currently faced with the challenge of increasing production to satisfy growing demand (Food and Agriculture Organisation (FAO), 2006), while meeting an international obligation to reduce greenhouse gas (GHG) emissions, which are the primary cause of global warming (Intergovernmental Panel on Climate Change (IPCC), 2007). This issue has led to an increasing interest in investigating and reducing the GHG emission intensity (kg of GHG per unit of product) of dairy production. In particular, the focus has been on improving the GHG emission intensity of primary dairy production, because the main sources of GHG from the dairy sector are emitted during this stage, for example, methane (CH4) from enteric fermentation (Hospido et al., Reference Hospido, Moreira and Feijoo2003; Steinfeld et al., Reference Steinfeld, Gerber, Wassenaar, Castel, Rosales and de Haan2006; Gerber et al., Reference Gerber, Vellinga, Opio, Henderson and Steinfeld2010).
An accepted method to assess the GHG emission intensity of primary dairy production is life cycle assessment (LCA; Thomassen et al., Reference Thomassen, van Calker, Smits, Iepema and de Boer2008b). LCA is a holistic systems approach that aims to quantify the potential environmental impacts (e.g. climate change, non-renewable energy use) generated throughout a product's life cycle, from raw-material acquisition through production, use, recycling and final disposal (International Organisation for Standardisation (ISO), 2006a). Thus, for primary dairy production, LCA quantifies GHG emissions from all processes associated with the dairy farm up to the point milk is sold from the farm (Cederberg and Mattsson, Reference Cederberg and Mattsson2000; Casey and Holden, Reference Casey and Holden2005). The general framework and requirements of LCA are internationally standardised (ISO, 2006a and 2006b). In addition, specific guidelines based on international standards have recently been developed for applying LCA to assess the GHG emission intensity of milk production (Carbon Trust, 2010; International Dairy Federation (IDF), 2010). However, LCA is not the recognised method for reporting the dairy sectors contribution to GHG emissions on a national basis. The standard method for reporting GHG is the IPCC guidelines (IPCC, 1996 and 2006). The IPCC method, unlike LCA, quantifies GHG emissions using a national sector-based approach (Schils et al., Reference Schils, Verhagen, Aarts and Šebek2005). The approach estimates emissions from the production and consumption of goods within defined national boundaries and emissions from the production of goods exported from a nation, but does not consider emissions from the production of goods imported into a country (Peters, Reference Peters2008).
The IPCC and LCA methods are frequently used to evaluate measures to reduce GHG emissions from dairy production. LCA studies of dairy systems have in particular assessed the effect of production system (e.g. conventional or organic dairy production) on GHG emissions (Cederberg and Mattsson, Reference Cederberg and Mattsson2000; Haas et al., Reference Haas, Wetterich and Köpke2001; Van der Werf et al., Reference Van der Werf, Kanyarushoki and Corson2009). In contrast, the IPCC approach has mainly been used to evaluate the effect that farm-management practices, such as nutrient management or feed additives, have on GHG emissions (Smith et al., Reference Smith, Martino, Cai, Gwary, Janzen, Kumar, McCarl, Ogle, O'Mara, Rice, Scholes, Sirotenko, Howden, McAllister, Pan, Romanenkov, Schneider, Towprayoon, Wattenbach and Smith2008). Few studies, to our knowledge, have applied the IPCC method to evaluate GHG emissions from different systems of production (Schils et al., Reference Schils, Verhagen, Aarts and Šebek2005), because the structure of the method is not conducive to a full appraisal of emissions from farming systems (Crosson et al., Reference Crosson, Shalloo, O'Brien, Lanigan, Foley, Boland and Kenny2011). However, the dairy sector can only achieve recognition for a reduction in GHG emissions when using the IPCC method.
The primary objective of this paper was to test the effect of applying the IPCC and LCA methodologies to compare GHG emissions from different systems of dairy production (conventional seasonal calving grass-based dairy system and a confinement dairy system). A secondary objective of this study was to determine what effect farm-management changes within both production systems would have on GHG emissions.
Material and methods
Dairy farm systems
The study quantified GHG emissions from a seasonal calving grass-based dairy system (Grass) and a confinement dairy system (Confinement) based on the work of Patton et al. (Reference Patton, Murphy and Butler2009) and Olmos et al. (Reference Olmos, Mee, Hanlon, Patton, Murphy and Boyle2009). Patton et al. (Reference Patton, Murphy and Butler2009) investigated the effect that grass and confinement dairy farms had on the annual milk performance of Holstein–Friesian cows, and Olmos et al. (Reference Olmos, Mee, Hanlon, Patton, Murphy and Boyle2009) examined the effect these dairy systems had on reproductive welfare.
The two studies were based on the same trial conducted at Moorepark Research Centre, Fermoy, Co. Cork, Ireland (52°16′N, 8°25′W) from December 2006 to January 2008 (Table 1). Spring-calving Holstein–Friesian cows were blocked based on genetic merit (Irish Economic Breeding Index €60, s.d. = €17.3), parity number (2.5, s.d. = 1.49), expected calving date (6 March, s.d. = 27.9 days), body condition score (3.0, s.d. = 0.69) and predicted annual milk yield (98 kg, s.d. = 123.4 kg) and assigned randomly from within pairs to either the Grass or Confinement system.
Technical description of the seasonal grass-based (Grass) and confinement dairy farms (Confinement) investigatedFootnote a
FPCM = fat and protein corrected milk; BW = body weight; LU = livestock unit, equivalent to one dairy cow; DM = dry matter.
a Data based on research of Patton et al. (Reference Patton, Murphy and Butler2009) and Olmos et al. (Reference Olmos, Mee, Hanlon, Patton, Murphy and Boyle2009).
b Off-farm land area required to produce purchased forage and concentrate. The land area was calculated using average yields from Central Statistics office (CSO, 2010) and Ecoinvent (2010).
c FPCM = (0.337 + 0.116 × %fat + 0.06 × %protein) × kg milk (Thomassen and de Boer, Reference Thomassen and de Boer2005).
d Forage intakes were estimated with the Moorepark Dairy System Model (Shalloo et al., Reference Shalloo, Dillon, Rath and Wallace2004) using milk production, animal BW, concentrate supplementation and feed-ration composition data from Patton et al. (Reference Patton, Murphy and Butler2009) and Olmos et al. (Reference Olmos, Mee, Hanlon, Patton, Murphy and Boyle2009).
Holstein–Friesian cows were housed full time in the Confinement system and fed total mixed ration (TMR). However, replacement heifers were only indoors during the winter period and grazed pasture during the growing season. Grass silage was harvested on-farm. The remaining feed was imported. Cows were offered specific TMR diets during the dry period (dry cow TMR) and lactation (lactation cow TMR; Table 2). The TMR diets offered were comparable in chemical composition to TMR fed in other studies (Kolver and Muller, Reference Kolver and Muller1998; Grainger et al., Reference Grainger, Auldist, O'Brien, Macmillan and Culley2009), apart from dry matter (DM) content, which was lower than that in other studies (Bargo et al., Reference Bargo, Muller, Delahoy and Cassidy2002; Vibart et al., Reference Vibart, Fellner, Burns, Huntington and Green2008). The low DM content was mainly due to the low DM of the maize and grass silages offered. The maize silage used was typical of maize produced in Ireland during a poor growing season with low DM content, poorly developed cobs and low starch content (Fitzgerald and Murphy, Reference Fitzgerald and Murphy1999).
Ingredient composition, DM and net energy of confinement dairy system (Confinement) dry and milking cow total mixed ration diets, seasonal grass-based dairy system (Grass) dry cow diet and concentrate offered to lactating cows at pasture
DM = total dry matter; UFL = unit feed lactation.
aVitamin A, 250 000 IU; vitamin D, 50 000 IU; vitamin E, 2000 IU; Se, 60 mg/kg; I, 700 mg/kg; Cu, 4000 mg/kg; Zn, 5000 mg/kg.
b1 UFL is 7.11 MJ of net energy.
The aim of the Grass system was to maximise the utilisation of grazed grass in the diet of lactating Holstein–Friesian dairy cows. This was accomplished by targeting calving to commence in early spring (mean calving date 14 February), which generally is the onset of the grass-growing season (Dillon et al., Reference Dillon, Crosse, Stakelum and Flynn1995). Grass silage was harvested on-farm when grass growth exceeded herd feed demand. The system was self-contained for forage. The N fertiliser input was 260 kg/ha and the total concentrate offered was 370 kg of DM per cow. Concentrate was offered mainly at the beginning and end of lactation. Grass silage was fed predominately during the dry period with minerals and vitamins included in the diet (Table 2).
The Moorepark Dairy System Model (MDSM; Shalloo et al., Reference Shalloo, Dillon, Rath and Wallace2004), a whole-farm simulation model, was used with milk production data, body weight (BW), concentrate supplementation data and the composition of the feed ration to predict the total dry matter intake (DMI) of animals. Data from Patton et al. (Reference Patton, Murphy and Butler2009) was used to check the accuracy of predicted feed intakes. The overall pregnancy rate was not significantly different between the two systems (Olmos et al., Reference Olmos, Mee, Hanlon, Patton, Murphy and Boyle2009); therefore, the replacement rate was the same for both systems (Table 1). Both farms investigated were specialised milk production systems, which meant that bull calves were sold from the farm as soon as practically possible.
GHG emissions
GHG emissions were calculated using a dairy farm GHG model (O'Brien et al., Reference O'Brien, Shalloo, Buckley, Horan, Grainger and Wallace2011). The model estimates the principal GHG emissions from primary dairy production: carbon dioxide (CO2), nitrous oxide (N2O) and CH4. The GHG model operates in combination with the MDSM (Shalloo et al., Reference Shalloo, Dillon, Rath and Wallace2004), which defines key input parameters (land area; animal inventory; milk production; feed intakes; herbage quality; grazing-season length; manure, fertiliser and lime application) required for the GHG model to quantify emissions.
In this study the model used two approaches, LCA and the IPCC method, to quantify GHG emissions from two dairy systems. The subsequent sections summarise how the LCA and IPCC methodologies were applied in the GHG model.
LCA methodology
The LCA approach was used to quantify emissions from all processes associated with the dairy production system up to the point that milk is sold from the farm. The approach was implemented in the GHG model by combining input data from the MDSM with specific LCA GHG emission factors in Microsoft Excel (Tables 3 and 4). The LCA emission factors for on-farm sources of CH4 and N2O emissions were obtained from the agricultural section of the Irish GHG national inventory (Duffy et al., Reference Duffy, Hyde, Hanley, Dore, O'Brien, Cotter and Black2011b). On-farm sources of CO2 were estimated using emission factors from the IPCC (2006) guidelines and data from Nemecek and Kägi (Reference Nemecek and Kägi2007). The LCA emission factors for off-farm GHG sources (GHG emissions associated with the production and supply of purchased farm inputs, e.g. synthetic fertiliser) were obtained mainly from European literature (Table 4) and databases included with the LCA software, SimaPro (Pré Consultants, 2008). The LCA approach did not consider GHG removals, because farm sinks (e.g. C sequestration) were assumed to be in equilibrium.
Emission factors used in the life cycle assessmentFootnote a approach for quantification of on-farm greenhouse gas emissions
CH4 = methane; DEI = digestible energy intake; S DMI = silage DMI; T DMI = total DMI; FL = feeding levels above maintenance energy; DMI = dry matter intake; GEI = gross energy intake; OM = organic matter; prop. = proportion; N2O-N = nitrous oxide – nitrogen; N = nitrogen; NO3 = nitrate; NH3 = ammonia; CO2 = carbon dioxide.
a All emissions associated with the dairy production system up to the point milk is sold from the farm.
b Replacements are dairy heifers aged 0 to 2 years.
c Dry matter content >20%.
d Manure excreted during grazing.
Emission factors used in the life cycle assessmentFootnote a approach for quantification of NH3-N emissions and off-farm greenhouse gas emissions
NH3-N = ammonia – nitrogen; N = nitrogen; luuk = UK livestock unit, equivalent to 500 kg BW; TAN = total ammoniacal nitrogen; CO2-eq = carbon dioxide equivalent; P2O5 = phosphorus pentoxide; DM = dry matter.
a All emissions associated with the dairy production system up to the point milk is sold from the farm.
b Dependent on age of animal.
c Dependent on slurry storage facility.
d DM content >20%.
e Yard = NH3-N from collecting yard prior to milking.
f Dependent on DM content of slurry and season of application.
g Synthetic N fertiliser applied to grazing area.
h Synthetic N fertiliser applied to silage, hay and cereal crops.
i Dependent on fertiliser compound and site of application (grassland or arable land).
j Concentrate blend used in the seasonal grass-based system.
k Concentrate blend used in the confinement system.
The LCA method allocates GHG emissions between multifunctional processes (O'Brien et al., Reference O'Brien, Shalloo, Buckley, Horan, Grainger and Wallace2011), including the production of concentrate co-products, such as barley grain and barley straw, and the production of milk and meat on-farm. The procedures used to allocate GHG emissions between co-products were updated using recently published LCA guidelines (IDF, 2010). For concentrate co-products, GHG emissions were allocated based on their relative economic values using data from Nemecek and Kägi (Reference Nemecek and Kägi2007), Central Statistics Office (CSO, 2010) and Ecoinvent (2010). Table 5 shows the composition, economic allocation and origin of concentrates used by the farming systems. For milk and meat, GHG emissions were allocated on the basis of the physiological feed requirements of the cow to produce milk and meat. The following equation from the IDF (2010) was used to calculate the allocation factor for milk and meat:
![\[--><$$>AF\, = \,1 - 5.7717\,\times \,({{M}_{{\rm{meat}}}}\,/\,{{M}_{{\rm{milk}}}})\eqno<$$><!--\]](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary:20160920214531914-0555:S1751731112000316:S1751731112000316_eqnU1.gif?pub-status=live){kind=small}
where AF is the allocation factor for milk, M meat the sum of BW of all animals including bull calves and culled mature animals and M milk the sum of mass of milk sold, corrected to 4% fat and 3.3% protein.
Concentrate ingredients for a seasonal grass-based dairy system (Grass) and a confinement dairy system (Confinement)
aThe economic allocation represents the percentage of emissions of the base product, for example, rapeseed allocated to the concentrate ingredient (e.g. rape meal).
bCitrus pulp substituted with beet pulp because of insufficient data.
GHG emissions calculated using LCA were estimated in terms of their 100-year global warming potentials (CO2 equivalents; CO2-eq), which on a weight basis relative to CO2 was set to a factor of 25 for 1 kg of CH4 and to a factor of 298 for 1 kg of N2O (Forster et al., Reference Forster, Ramaswamy, Artaxo, Berntsen, Betts, Fahey, Haywood, Lean, Lowe, Myhre, Nganga, Prinn, Raga, Schulz and van Dorland2007). The temporal coverage of the LCA method in the GHG model was a period of 1 year. Thus, the approach generated a static account of dairy systems’ annual on-farm, off-farm and total (on-farm plus off-farm) GHG emissions. Dairy systems produce food and provide environmental services (e.g. landscape conservation). Consequently, LCA estimates of annual GHG emissions were expressed per unit of farm area (t CO2-eq/on-farm ha) and per unit of product (kg CO2-eq/kg FP (fat and protein)). Furthermore, LCA estimates of annual GHG emissions per unit area and per unit of product were updated also to estimate emissions per total ha (on-farm and off-farm hectares required to produce purchased forage and concentrate) and per kg of fat and protein corrected milk (FPCM).
IPCC methodology
The sources of GHG considered using the IPCC method are enteric fermentation, manure management and agricultural soils (emissions of N2O from synthetic and organic fertiliser use, manure excreted by grazing cattle and N2O emissions from ammonia (NH3) re-deposition and nitrate (NO3) leaching). These are the sources the IPCC consider relevant for dairy farming (Schils et al., Reference Schils, Verhagen, Aarts and Šebek2005). The IPCC approach in the GHG model excludes GHG removals from farm sinks, because GHG sinks are not reported in the agricultural sector. The IPCC method was applied to quantify GHG emissions from dairy cows and replacement heifers (0 to 2 years).
The IPCC method was implemented in the GHG model using IPCC emission factors from the agriculture section of the Irish GHG national inventory (Duffy et al., Reference Duffy, Hyde, Hanley, Dore, O'Brien, Cotter and Black2011b; Table 6) and the parameters generated by the MDSM. The IPCC emission factors are categorised into three different tiers, with each successive tier having an increased level of detail and accuracy. Presently, the Irish GHG inventory uses IPCC emission factors from different tiers to calculate emissions from dairy production. Thus, the IPCC method in the GHG model uses tier-1 emission factors to calculate N2O emissions from agricultural soils and tier-2 emission factors to estimate CH4 emissions from enteric fermentation and manure management (Duffy et al., Reference Duffy, Hyde, Hanley, Dore, O'Brien, Cotter and Black2011b). GHG emissions quantified using the IPCC method were estimated in terms of their 100-year global warming potential using the same procedure as described for the LCA method.
IPCCFootnote a emission factors for quantification of dairy farm GHG emissions
IPCC = Intergovernmental Panel on Climate Change; GHG = greenhouse gas; CH4 = methane; DEI = digestible energy intake; S DMI = silage DMI; T DMI = total DMI; FL = feeding levels above maintenance energy; DMI = dry matter intake; GEI = gross energy intake; OM = organic matter; prop. = proportion; N2O = nitrous oxide; N = nitrogen; NH3 = ammonia; NO2 = nitrate.
a IPCC emission factors as applied in Ireland for quantification of GHG emissions (Duffy et al., Reference Duffy, Hyde, Hanley, Dore, O'Brien, Cotter and Black2011b).
b Replacements are dairy heifers aged 0 to 2 years.
c DM content >20%.
d Manure excreted during grazing.
The IPCC method in the GHG model produces a static account of annual GHG emissions from dairy production. Annual GHG emissions quantified using the IPCC method were expressed per unit of product to test the effect of methodology and per on-farm ha, because national inventories of GHG emissions are assessed in isolation from food production, for example, annual GHG emissions per nation state (Casey and Holden, Reference Casey and Holden2005; Crosson et al., Reference Crosson, Shalloo, O'Brien, Lanigan, Foley, Boland and Kenny2011). GHG emissions per unit of product were allocated between milk and meat using the IDF (2010) approach previously described for the LCA method.
Scenario analysis
Farm-management changes to reduce calculated GHG emissions, using IPCC and LCA methodologies, were simulated with the GHG model for both dairy systems using scenario modelling. The following scenarios were simulated:
-
• increasing forage quality to achieve a higher net energy content (+0.05 unit feed lactation (UFL)/kg DM; Jarrige, Reference Jarrige1989);
-
• increasing milk production without compromising fertility and other important traits via genetic improvement (+500 kg milk produced per cow);
-
• reducing on-farm synthetic N fertiliser application to decrease the farm N surplus (−25 kg N/ha);
-
• reducing the replacement rate (−3%) to reduce the proportion of non-milk producing animals in the herd;
-
• covering the manure store and flaring of CH4.
The management changes were simulated one by one and in combination to determine the cumulative efficacy of the selected farm-management changes, which were not all directly additive. The present IPCC guidelines were used to predict GHG emissions from enclosed manure stores where gas is captured (IPCC, 2006). Scenario modelling also was used to test the effects of C sequestration by grassland soils on dairy system GHG emissions. Carbon sequestration by grassland soils is caused predominately by farm-management changes, for instance the conversion of arable land to grassland (Soussana et al., Reference Soussana, Tallec and Blanfort2010). Thus, the scenario analysis assumed that the grassland area of both dairy systems was newly established (<2 years) and had previously been used for growing arable crops. The simulated rate of sequestration of C by grassland soils was estimated as 2.5 t C/ha per year based on the review of Soussana et al. (Reference Soussana, Tallec and Blanfort2010). The ratio of molecular weights of CO2 and C were used to quantify CO2 sequestration (3.67 kg CO2/kg C).
Results
GHG emissions of grass and confinement dairy systems
The GHG emissions of the Grass system and the Confinement system, calculated using LCA, are shown in Tables 7 and 8, respectively. GHG emissions quantified using the IPCC method for both systems are shown in Table 9. The LCA results show that the Confinement system decreased on-farm GHG emissions (kg CO2-eq) per tonne of product (FPCM and FP) on average by 9% relative to the Grass system (Tables 7 and 8). However, when off-farm GHG emissions were included, the Confinement system increased total GHG emissions per unit of product by 16% compared with the Grass system. The IPCC results (Table 9) per tonne of product show that GHG emissions were, on average, 8% lower for the Confinement system than the Grass system.
Effect of selected farm-management changes and C sequestration on the annual GHG emissions of a seasonal grass-based dairy system (Grass) calculated using life cycle assessment with a GHG model (O'Brien et al., Reference O'Brien, Shalloo, Buckley, Horan, Grainger and Wallace2011)
C sequestration = carbon sequestration; GHG = greenhouse gas; N = nitrogen; CH4 = methane; N2O = nitrous oxide; NO3 = nitrate; NH3 = ammonia; CO2 = carbon dioxide; UFL = unit feed lactation; DM = dry matter; LU = livestock unit equivalent to one dairy cow; FP = Kg of milk fat plus protein; FPCM = fat and protein corrected milk.
aMilk production per cow of Grass system increased while maintaining fertility performance.
bN surplus per ha of Grass system reduced by 25 kg N/ha via reduction in synthetic N use.
cReplacement rate of Grass system reduced from 18% to 15%.
dManure stores of Grass system covered and CH4 flared.
ePredicted cumulative efficacy of selected farm-management changes.
fPermanent grassland C sequestration increased from 0 to 2.5 t C/ha per year.
g1 UFL is 7.11 MJ of net energy.
hCO2-eq = CO2 equivalent where CO2 = 1; CH4 = 25; NO2 = 298.
iProduction of fuel, electricity and lime.
jManure excreted on pasture by grazing cattle.
kTotal = on-farm GHG emissions or ha and off-farm GHG emissions or ha.
lFPCM = (0.337 + 0.116 × %fat + 0.06 × %protein) × kg milk (Thomassen and de Boer, Reference Thomassen and de Boer2005).
Effect of selected farm-management changes and C sequestration on the annual GHG emissions of a confinement dairy system (Confinement) calculated using life cycle assessment with a GHG model (O'Brien et al., Reference O'Brien, Shalloo, Buckley, Horan, Grainger and Wallace2011)
C sequestration = carbon sequestration; GHG = greenhouse gas; N = nitrogen; CH4 = methane; N2O = nitrous oxide; NO3 = nitrate; NH3 = ammonia; CO2 = carbon dioxide; UFL = unit feed lactation; DM = dry matter; LU = livestock unit equivalent to one dairy cow; FP = Kg of milk fat plus protein; FPCM = fat and protein corrected milk.
aMilk production per cow of Confinement system increased while maintaining fertility performance.
bN surplus per ha of Confinement system reduced by 25 kg N/ha via reduction in synthetic N use.
cReplacement rate of Confinement system reduced from 18% to 15%.
dManure stores of Confinement system covered and CH4 flared.
ePredicted cumulative efficacy of selected farm-management changes.
fPermanent grassland C sequestration increased from 0 to 2.5 t C/ha per year.
g1 UFL is 7.11 MJ of net energy.
hCO2-eq = CO2 equivalent where CO2 = 1; CH4 = 25; NO2 = 298.
iProduction of fuel, electricity and lime.
jManure excreted on pasture by grazing cattle.
kTotal = on-farm GHG emissions or ha and off-farm GHG emissions or ha.
lFPCM = (0.337 + 0.116 × %fat + 0.06 × %protein) × kg milk (Thomassen and de Boer, Reference Thomassen and de Boer2005).
Effect of selected farm-management changes and C sequestration on the annual GHG emissions of a seasonal grass-based dairy system (Grass) and a confinement dairy system (Confinement) calculated using the Intergovernmental Panel on Climate Change method with a GHG model (O'Brien et al., Reference O'Brien, Shalloo, Buckley, Horan, Grainger and Wallace2011)
C sequestration = carbon sequestration; GHG = greenhouse gas; N = nitrogen; CH4 = methane; CO2 = carbon dioxide; N2O = nitrous oxide; NO3 = nitrate; NH3 = ammonia; UFL = unit feed lactation; DM = dry matter; LU = livestock unit equivalent to one dairy cow; FP = Kg of milk fat plus protein; FPCM = fat and protein corrected milk.
aMilk production per cow of baseline farms increased while maintaining fertility performance.
bN surplus per ha of baseline farms reduced by 25 kg N/ha via reduction in synthetic N use.
cReplacement rate of baseline farms reduced from 18% to 15%.
dManure stores of baseline farms covered and CH4 flared.
ePredicted cumulative efficacy of selected farm-management changes.
f1 UFL is 7.11 MJ of net energy.
gCO2-eq = CO2 equivalent where CO2 = 1; CH4 = 25; NO2 = 298.
hManure excreted on pasture by grazing cattle.
ijFPCM = (0.337 + 0.116 × %fat + 0.06 × %protein) × kg milk (Thomassen and de Boer, Reference Thomassen and de Boer2005).
The LCA approach found that on-farm GHG emissions per on-farm ha were higher for the Confinement system than the Grass system (Tables 7 and 8). However, per total (on and off-farm area) ha, total GHG emissions were 11% greater for the Grass system than the Confinement system. Per on-farm ha, the IPCC results show that GHG emissions were lower for the Grass system than the Confinement system (Table 9).
GHG emissions components
LCA
Methane accounted for 51% of total GHG emissions (total of CO2-eq on-farm and off-farm) in the Grass system and for 57% in the Confinement system, N2O accounted for 36% in the Grass system and for 15% in the Confinement system and CO2 accounted for 13% in the Grass system and for 27% in the Confinement system (Tables 7 and 8). The main contributors to total GHG emissions in the Grass system were CH4 emissions from enteric fermentation (44%), N2O emissions from excreta deposited during grazing (13%); GHG emissions associated with the production of fertiliser (10%), and N2O emissions from synthetic fertiliser application (9%; Table 7). Enteric fermentation was the greatest contributor to on-farm GHG emissions (54%) in the Grass system followed by N2O from excreta deposited during grazing (16%), N2O from fertiliser application (11%) and CH4 emissions from slurry storage (9%). Enteric CH4 emissions per unit of intake for dairy cows in the Grass system were 23.1 g/kg DMI. Off-farm GHG emissions consisted mainly of GHG emissions from the production of fertiliser (57%) and concentrate feed (26%; Table 7).
The main contributors to total GHG emissions in the Confinement system were CH4 emissions from enteric fermentation (36%), GHG emissions from the production of concentrate (25%) and CH4 emissions from manure storage (20%; Table 8). On-farm GHG emissions in the Confinement system were mainly due to enteric CH4 (57%), but CH4 losses per unit of intake for dairy cows were 4% lower for the Confinement system (22.3 g/kg DMI) than the Grass system. Methane from manure storage was the next largest source of on-farm GHG emissions (31%). The production of concentrate feed was the main source of off-farm GHG emissions (69%) followed by CO2 emissions from electricity, fuel and lime production (15%; Table 8). The main sources of GHG emissions from concentrate production were CO2 emissions from the conversion of rainforest to agricultural land (35%), GHG emissions from energy and fertiliser production (17%), CO2 emissions from agricultural field operations and transport of feed (15%) and N2O emissions from fertiliser application (12%).
IPCC method
Methane accounted for 64% of GHG emissions (in CO2-eq) in the Grass system and for 90% in the Confinement system, N2O accounted for 36% in the Grass system and for 10% in the Confinement system (Table 9). The main sources of GHG emissions in the Grass system were CH4 emissions from enteric fermentation (55%) and manure storage (9%), N2O emissions from manure excreted by grazing cattle (16%), N2O from synthetic fertiliser application (11%) and N2O from NO3 − leaching and NH3 re-deposition (6%; Table 9). Methane from enteric fermentation (58%) and N2O and CH4 from manure storage (33%) were the main contributors to GHG emissions in the Confinement system (Table 9).
Scenario analysis
LCA
The predicted effects of selected farm-management changes and C sequestration on GHG emissions from the Grass and Confinement systems calculated using the LCA method are shown in Tables 7 and 8.
Improving forage quality reduced on-farm and total GHG emissions per unit of product and area by 4% to 5% for the Grass system (Table 7) and by 3% for the Confinement system (Table 8). The strategy reduced CH4 emissions from enteric fermentation and manure storage, N2O emissions from fertiliser use and manure excreted by grazing animals and off-farm emissions from concentrate and fertiliser production (Tables 7 and 8).
Increasing milk production through genetic improvement increased GHG emissions per livestock unit (LU; equivalent to one dairy cow) of both dairy systems (3%). However, the increase in milk production (6% to 8%) was greater than the increase in dairy systems’ GHG emissions. Therefore, GHG emissions per unit of product decreased (3% to 4%). Increasing milk production had little effect (<0.7%) on GHG emissions per unit area, because the on- and off-farm area of both dairy systems increased (3% to 4%) to supply the extra feed required for the increased milk production.
Reducing the farm N surplus reduced the on-farm and total GHG emissions per unit of product and area for both systems (1% to 2%; Tables 7 and 8). The strategy reduced N2O emissions from on-farm fertiliser use, N2O from NO3 − leaching and NH3 re-deposition and off-farm emissions from fertiliser production.
For both dairy systems, the LCA method found that reducing the replacement rate increased GHG emissions per LU (Tables 7 and 8), because the ratio of dairy cows to replacement animals increased. However, GHG emissions per unit area decreased (1% to 2%), because the number of replacement animals and the stocking rate declined. Reducing the replacement rate had little effect on GHG emissions per unit of product (Tables 7 and 8), because reducing the replacement rate decreased GHG emissions per unit area but increased the percentage of GHG emissions allocated to milk rather than to beef (Tables 7 and 8).
Covering the manure store and flaring CH4 reduced on-farm and total GHG emissions per unit of product and area from the Grass system by 6% and 5%, respectively (Table 7). On-farm and total GHG emissions per unit of product and area decreased by 22% and 14%, respectively, when CH4 from manure storage was flared in the Confinement system (Table 8). The management change eliminated N2O emissions from slurry storage and significantly reduced CH4 emissions from slurry storage.
Combining farm-management changes reduced on-farm GHG emissions per unit of product by 15% for the Grass system and by 28% for the Confinement system (Tables 7 and 8). Per on-farm ha, combining management changes reduced on-farm GHG emissions by 8% for the Grass system and by 22% for the Confinement system. When off-farm emissions were included, combining management changes reduced total GHG emissions per unit of product by 15% for the Grass system and 20% for the Confinement system. Per total ha, combining management changes reduced total GHG emissions by 7% for the Grass system and by 15% for the Confinement system (Tables 7 and 8).
Assuming that the grassland area of both dairy systems was newly established (<2 years) and had previously been used for growing arable crops reduced on-farm GHG emissions per unit of product and area by 30% for the Confinement system and by 65% for the Grass system. Similarly, the assumption reduced total GHG emissions per unit of product and area by 19% for the Confinement system and by 53% for the Grass system. The reductions were caused by grassland C sequestration (Tables 7 and 8).
IPCC method
The predicted effect of selected farm-management changes on GHG emissions from the Grass and Confinement systems calculated using the IPCC method are shown in Table 9. The IPCC method results show that in the Grass and Confinement systems’ GHG emissions per unit of product and area were reduced by improving forage quality (3% to 5%), reducing the farm N surplus (1%) and capturing CH4 from manure stores (6% to 24%). Increasing milk production per cow reduced baseline dairy systems’ GHG emissions per unit of product (3% to 4%), but had a minor effect on GHG emissions per unit area (<0.7%). Reducing the replacement rate decreased baseline dairy systems’ GHG emissions per unit area (2%), because the stocking rate declined. However, the strategy had little effect on GHG emissions per unit of product, because reducing the replacement rate increased the proportion of GHG emissions allocated to milk compared with beef (Table 9). The IPCC approach found that combining management changes reduced GHG emissions per unit of product by 16% for the Grass system and by 30% for the Confinement system (Table 9). Per unit area, combining management changes reduced GHG emissions by 9% for the Grass system and by 24% for the Confinement system (Table 9).
Discussion
Studies that have compared GHG emissions from seasonal calving grass-based and confinement dairy systems generally have used environmental system analysis approaches such as LCA (Arsenault et al., Reference Arsenault, Tyedmers and Fredeen2009; Rotz et al., Reference Rotz, Soder, Skinner, Dell, Kleinman, Schmidt and Bryant2009). The IPCC method of quantifying agricultural GHG emissions has rarely been used to assess emissions from these different systems of dairy production (Schils et al., Reference Schils, Verhagen, Aarts, Kuikman and Sebek2006; Browne et al., Reference Browne, Eckard, Behrendt and Kingwell2011). Thus, the application of the LCA and IPCC methods in this study provides a unique opportunity to determine if these methods agree on their predictions of GHG emissions from seasonal grass-based and confinement dairy systems. However, given that our comparison of dairy systems is based on a single farm for each system, it cannot be considered to be a representative comparison of grass and confinement systems of dairy production.
Methodology comparison of grass-based and confinement dairy systems’ GHG emissions: IPCC v. LCA
Effect of dairy systems on GHG emissions
Consistent with previous international studies, the LCA and IPCC methods agreed that CH4 from enteric fermentation was the main cause of on-farm GHG emissions for seasonal grass-based and confinement dairy systems (Schils et al., Reference Schils, Verhagen, Aarts and Šebek2005 and Reference Schils, Verhagen, Aarts, Kuikman and Sebek2006; Rotz et al., Reference Rotz, Soder, Skinner, Dell, Kleinman, Schmidt and Bryant2009; O'Brien et al., Reference O'Brien, Shalloo, Buckley, Horan, Grainger and Wallace2011). Both methods also agreed with the previous studies that the other main sources of on-farm GHG emissions for the grass-based system were N2O emissions from manure excreted by grazing cattle and N2O emissions from fertiliser application (Schils et al., Reference Schils, Verhagen, Aarts, Kuikman and Sebek2006; Basset-Mens et al., Reference Basset-Mens, Kelliher, Ledgard and Cox2009). Similarly, the IPCC and LCA methods supported previous analysis that on-farm GHG emissions in the confinement system were emitted mainly from manure storage when enteric CH4 emissions were excluded (Arsenault et al., Reference Arsenault, Tyedmers and Fredeen2009; Rotz et al., Reference Rotz, Montes and Chianese2010). Thus, apart from enteric CH4 emissions, the main sources of on-farm GHG emissions differed between dairy systems. The contribution of on-farm sources to GHG emissions varied between systems, because of differences in the housing period of dairy cows and the proportion of feed produced on-farm. Therefore, system specific strategies should be developed to achieve the largest reduction in on-farm GHG emissions from grass-based and confinement dairy systems.
Several studies have reported that feed intake is the principal determinant of enteric CH4 emissions from cattle (Mills et al., Reference Mills, Kebreab, Yates, Crompton, Cammell, Dhanoa, Agnew and France2003; Hegarty et al., Reference Hegarty, Goopy, Herd and McCorkell2007; O'Neill et al., Reference O'Neill, Deighton, O'Loughlin, Mulligan, Boland, O'Donovan and Lewis2011). In addition, feed intake is also a key variable required to quantify CH4 and N2O emissions from manure (Basset-Mens et al., Reference Basset-Mens, Kelliher, Ledgard and Cox2009). Therefore, both methods showed, similar to previous results, that on-farm GHG emissions per LU was greater (8% to 9%) for the confinement system than the grass-based system, because grass-fed cows produced 18% less milk per LU and thus consumed 16% less feed (Capper et al., Reference Capper, Cady and Bauman2009; Rotz et al., Reference Rotz, Soder, Skinner, Dell, Kleinman, Schmidt and Bryant2009). Feed intake per unit area, which is predominately controlled by the stocking rate of the dairy system, is also a major determinant of GHG emissions per unit area. Thus, decreasing the on-farm stocking rate generally reduces GHG emissions per on-farm ha (Casey and Holden, Reference Casey and Holden2005).
Our results, irrespective of method, supported this relationship and showed that the dairy system with the higher on-farm stocking rate, the confinement system (5 LU/on-farm ha), emitted twice as much GHG emissions per on-farm ha than the lower stocked grass-based system (2.5 LU/on-farm ha). However, per unit of product, both methods observed that on-farm GHG emissions were greater for the grass-based system than for the confinement system. This was because TMR-fed Holstein–Friesian cows in the confinement system produced ∼2.4 times more milk per on-farm ha and 21% more milk per LU than forage-fed cows (Kolver and Muller, Reference Kolver and Muller1998; Bargo et al., Reference Bargo, Muller, Delahoy and Cassidy2002). In addition, TMR-fed cows emitted less enteric CH4 as a proportion of total DMI (Johnson and Johnson, Reference Johnson and Johnson1995; Bell et al., Reference Bell, Wall, Russell, Morgan and Simm2010).
The LCA method also quantifies off-farm GHG emissions related to dairy systems. Similar to previous dairy LCA studies, CO2 and N2O emissions from the cultivation and transport of imported feed and GHG emissions from the manufacture of synthetic N fertiliser were the main contributors to off-farm GHG emissions (Thomassen et al., Reference Thomassen, van Calker, Smits, Iepema and de Boer2008b; Van der Werf et al., Reference Van der Werf, Kanyarushoki and Corson2009). Off-farm GHG emissions from the grass-based system were less than half of those from the confinement system, because the confinement system consumed more imported feedstuffs than the grass-based system. As a result of the large difference between dairy systems’ off-farm GHG emissions, the LCA approach found that the increase in milk yield per LU moving from the grass-based system to the confinement system (21%) was less than the increase in total GHG emission per LU (38%).
Thus, similar to previous studies, total GHG emissions per unit of product and LU were higher for the confinement system than the grass-based system (Rotz et al., Reference Rotz, Soder, Skinner, Dell, Kleinman, Schmidt and Bryant2009; Flysjö et al., Reference Flysjö, Henriksson, Cederberg, Ledgard and Englund2011). However, per total (on-farm and off-farm) ha, total GHG emissions were lower for the confinement system than the grass-based system, because the stocking rate per total ha of the confinement system was lower than the grass-based system. The stocking rate per total ha of the confinement system was lower than the grass system due to the large area required for the off-farm production of key components of TMR such as soya bean meal (3.2 m2/kg DM). Therefore, this finding also supports the results of Casey and Holden (Reference Casey and Holden2005) that reducing stocking rate decreases GHG emissions per ha.
Effect of boundary on dairy systems’ GHG emissions
The IPCC method, unlike the LCA approach, does not attribute GHG emissions from the production of farm imports to primary dairy production, because these emissions are outside national boundaries or ascribed to another sector, such as energy (Schils et al., Reference Schils, Verhagen, Aarts and Šebek2005). Furthermore, the approach does not attribute the off-farm area required to produce farm imports to dairy systems. Omitting off-farm GHG sources and the off-farm area associated with dairy production would not matter if farming systems always ranked the same whether the IPCC or LCA methods were used. For example, our results show that excluding off-farm GHG sources had no effect on the ranking of dairy systems’ GHG emissions per LU. However, when GHG emissions were expressed per unit of product, the IPCC method indicated that the grass system had higher emissions, but the LCA approach showed that the grass system had lower emissions. The different ranking of dairy systems’ GHG emissions per unit of product using the IPCC and LCA method agrees with previous work and can be explained by the exclusion of emissions associated with upstream production chains (e.g. concentrate production) using the IPCC method (O'Brien et al., Reference O'Brien, Shalloo, Buckley, Horan, Grainger and Wallace2011).
Per unit area, the exclusion of off-farm area required for farm imports also affected the rankings of dairy systems’ GHG emissions. The IPCC method showed that per on-farm ha the grass-based system emitted lower GHG emissions, because the stocking rate was 50% lower than the confinement system. However, per total ha, the stocking rate of the confinement system was lower than the grass-based system, because the confinement system was more reliant on imported feed and thus required a greater off-farm area (Table 1). Consequently, the LCA method found that dairy systems had the opposite effect on GHG emissions per total ha. Therefore, the methods disagreed, because including the off-farm area required for farm imports affected the stocking rate of dairy systems, which is a key driver of GHG emissions per unit area (Casey and Holden, Reference Casey and Holden2005).
The disagreement between the IPCC and LCA methods per unit of product and area highlights that it is incorrect to only consider on-farm components of the dairy system relevant for policy reporting (Schils et al., Reference Schils, Verhagen, Aarts, Kuikman and Sebek2006; O'Brien et al., Reference O'Brien, Shalloo, Grainger, Buckley, Horan and Wallace2010 and Reference O'Brien, Shalloo, Buckley, Horan, Grainger and Wallace2011). Instead, holistic approaches such as LCA give a more realistic assessment of GHG emissions from differing dairy production systems. Holistic approaches increase emissions attributed to dairy systems relative to approaches that consider only part of the production system, but are more likely to ensure that GHG emissions are reduced and not transferred to another location (e.g. ‘carbon leakage’) or part of the production system. The inadequacy of the IPCC method for assessing GHG emissions from different dairy systems does not mean the approach should no longer be used to report national GHG emissions (Schils et al., Reference Schils, Verhagen, Aarts, Kuikman and Sebek2006; Crosson et al., Reference Crosson, Shalloo, O'Brien, Lanigan, Foley, Boland and Kenny2011). However, to achieve a reduction in GHG emissions, reform of the IPCC method is needed given the problems identified in our study and previous studies (Andrew and Forgie, Reference Andrew and Forgie2008; Peters, Reference Peters2008; Peters and Hertwich, Reference Peters and Hertwich2008).
Consequently, it has been suggested that countries should quantify GHG emissions associated with national consumption (production and imports minus exports) in addition to estimating GHG emissions produced within a nation (Peters, Reference Peters2008; Peters and Hertwich, Reference Peters and Hertwich2008). This approach would overcome some of the problems with the present IPCC method, such as carbon leakage and credit dairy practices that reduce emissions in other sectors or nations. However, consumption-based national GHG accounting would require political decisions to extend outside national boundaries, which is a significant barrier to the implementation of the approach (Peters, Reference Peters2008). Furthermore, consumption-based national GHG inventories require more complex calculations and hence assumptions, thereby increasing uncertainty and reducing the accuracy of GHG national inventories (Peters, Reference Peters2008).
Effect of unit of expression on GHG emissions
The IPCC and LCA methods use different approaches to report GHG emissions. The IPCC guidelines report annual GHG emissions per nation state (Crosson et al., Reference Crosson, Shalloo, O'Brien, Lanigan, Foley, Boland and Kenny2011). Thus, dairy farm GHG emissions are frequently assessed in isolation from dairy production for instance, per unit area or per LU (Casey and Holden, Reference Casey and Holden2005; Capper et al., Reference Capper, Cady and Bauman2009). In contrast, the LCA approach is typically applied to quantify GHG emissions per unit of product for dairy production systems, for example, Arsenault et al. (Reference Arsenault, Tyedmers and Fredeen2009) and Flysjö et al. (Reference Flysjö, Henriksson, Cederberg, Ledgard and Englund2011). Previous studies have shown that expressing GHG emissions per unit area, per LU or per unit of product can change the ranking of dairy systems’ GHG emissions (Casey and Holden, Reference Casey and Holden2005; Capper et al., Reference Capper, Cady and Bauman2009; O'Brien et al., Reference O'Brien, Shalloo, Buckley, Horan, Grainger and Wallace2011). Our results agreed with these findings; for instance, the IPCC approach found that the confinement system had lower emissions per unit of product and higher emissions per on-farm ha and per LU than the grass-based system. Similarly, the LCA method showed that the confinement system emitted lower total emissions per total ha than the grass-based system, but per unit of product the dairy systems had the opposite effect.
Thus, these findings show that reducing emissions on a livestock or area basis will not lead to the highest reduction in GHG emissions possible for a given level of product. The FAO (2006) predict that between 2000 and 2050 demand for dairy products will double due to a growing affluent human population. Therefore, if dairy production is to decrease emissions and supply sufficient milk to satisfy future demand, GHG emissions should be quantified per unit of product (Capper et al., Reference Capper, Cady and Bauman2009; del Prado et al., Reference del Prado, Chadwick, Cardenas, Misselbrook, Scholefield and Merino2010). Expressing GHG emissions on a product basis would incentivise producers to adopt practices to increase the efficiency of production, which reduces GHG emissions for a given level of production (Beukes et al., Reference Beukes, Gregorini, Romera, Levy and Waghorn2010). Furthermore, it would encourage the production of milk production within specific regions where it can be produced with the least impact on GHG emissions. In contrast, mitigation strategies that reduce emissions per LU or per unit area may not reduce emissions per unit of product, because assessing emissions on a livestock or area basis does not assess the effect mitigation strategies can have on milk production.
Effect of selected farm-management changes on dairy systems’ GHG emissions
The IPCC and LCA methods were also used in this study to determine, via scenario analysis, if farm-management changes could be introduced to reduce GHG emissions from grass-based and confinement dairy systems. The results indicated, similar to previous studies, that single management changes to improve animal performance, increase N-use efficiency or improve forage quality can reduce GHG emissions by improving productive efficiency (Schils et al., Reference Schils, Verhagen, Aarts and Šebek2005; Lovett et al., Reference Lovett, Shalloo, Dillon and O'Mara2008; Beukes et al., Reference Beukes, Gregorini, Romera, Levy and Waghorn2010). Similarly, our findings agreed with previous studies that individual strategies to control emissions from manure storage, in particular flaring CH4 emissions, substantially reduce on-farm (22% to 24%) and total (14%) GHG emissions from confinement systems (Weiske et al., Reference Weiske, Vabitsch, Olesen, Schelde, Michel, Friedrich and Kaltschmitt2006; Rotz et al., Reference Rotz, Montes and Chianese2010). However, for grass-based dairy producers, the results of this study suggest that reducing emissions from manure storage does not have a large effect on GHG emissions (5% to 7%), because most manure is voided during grazing. As expected, the ability of strategies to reduce GHG emissions was affected by changes in off-farm GHG emissions. However, the effect of management changes on off-farm emissions could only be considered when using the LCA approach. Thus, our study demonstrated, similar to Schils et al. (Reference Schils, Verhagen, Aarts and Šebek2005) that the IPCC method is not suitable for analysis of mitigation strategies, because it does not assess the effect that farm-management changes have on off-farm GHG emissions.
Although individual management changes generally had a small effect on GHG emissions (<5%), when several management strategies were combined the on-farm (15% to 30%) and total (15% to 20%) GHG emission intensity of dairy systems was reduced significantly. Thus, these results suggest that dairy producers can play a major role in addressing climate change if the benefits associated with the mitigation strategy can be credited to the producer. However, GHG mitigation strategies will be successfully adopted by farmers only if they are also shown to be profitable. Assessing the cost-effectiveness of farm-management changes was outside the scope of this study, but previous reports indicate that most management changes tested in this study increase the profitability of dairy production (Lovett et al., Reference Lovett, Shalloo, Dillon and O'Mara2008; Beukes et al., Reference Beukes, Gregorini, Romera, Levy and Waghorn2010). Indeed, the only strategy evaluated in this study that previous reports indicate is not economically viable was the flaring of CH4 from manure storage (Weiske et al., Reference Weiske, Vabitsch, Olesen, Schelde, Michel, Friedrich and Kaltschmitt2006). However, considering the rapid increase in energy prices in recent years and potential market subsidies, capturing and flaring CH4 from manure stores for electricity generation or heating may prove to be a viable option in the future.
The analysis of the effect of management changes on dairy systems’ GHG emissions also revealed that the procedure used to allocate emissions between co-products can influence LCA results. For instance, the results of this study demonstrate that when all GHG emissions were attributed to milk, reducing the annual cow replacement rate reduces emissions from non-milk producing animals and therefore GHG emissions per unit of product. This finding is consistent with previous studies that used this allocation technique (Beukes et al., Reference Beukes, Gregorini, Romera, Levy and Waghorn2010; Vellinga et al., Reference Vellinga, de Haan, Schils, Evers and van den Pol-van Dasselaar2011). However, when emissions were allocated between milk and meat based on the physiological feed requirements of the animal to produce these products, the management strategy had no effect. The strategy of reducing replacement rate had no effect using the allocation technique used in this study, because the proportion of emissions allocated to milk increased, which offset the reduction in GHG emissions from non-milk producing animals. The large influence that different allocation procedures have on LCA results supports the most recent international standards recommendation that LCA studies should avoid allocation by expanding the boundaries of the production system to include the avoided product (e.g. meat; ISO, 2006a). This approach increases the complexity of LCA studies and requires more data but overcomes the allocation problem identified in this and previous studies (Cederberg and Stadig, Reference Cederberg and Stadig2003; Thomassen et al., Reference Thomassen, Dalgaard, Heijungs and de Boer2008a).
Carbon Sequestration
Another important issue in the evaluation of GHG emissions from dairy systems is C sequestration by grassland soils. However, studies often do not consider or mention C sequestration (e.g. Casey and Holden, Reference Casey and Holden2005). Studies that have considered C sequestration by grassland soils have reported that the process makes a substantial (37% to 77%), albeit short-lived contribution (20 to 50 years) to reducing GHG emissions from dairy systems (Schils et al., Reference Schils, Verhagen, Aarts and Šebek2005; Rotz et al., Reference Rotz, Soder, Skinner, Dell, Kleinman, Schmidt and Bryant2009). Our scenario analysis findings for the LCA method agree that C sequestration following the conversion of arable land to grassland can substantially reduce GHG emissions, but the results of the IPCC approach did not support these findings. This was because the IPCC method does not report C sequestration by grassland soils in the agricultural sector (Schils et al., Reference Schils, Verhagen, Aarts and Šebek2005). Thus, until the IPCC method is reformed, farm-management practices that improve C sequestration will not reduce reported GHG emissions from the agricultural sector.
Conclusion
The results of the IPCC and LCA methods show that reducing dairy systems’ GHG emissions on a per unit area basis is not necessarily the most effective approach to reduce GHG emissions for a given level of product. Therefore, given the growing demand for dairy products, mitigation efforts should focus on reducing GHG emissions per unit of product. The IPCC and LCA methods did not rank grass-based and confinement dairy systems’ emissions per unit of product in the same order, because the IPCC method excludes GHG emissions from farm supply chains (e.g. concentrate production). Thus, a complete assessment of GHG emissions per unit of product for differing dairy systems can be made only using holistic approaches such as LCA. Scenario analysis showed that GHG emissions per unit of product of both dairy systems could be reduced significantly when several farm-management practice changes were combined. In addition, the analysis showed that the LCA method should be used to assess changes in GHG emissions per unit of product. However, further standardisation of the LCA methodology is needed given the sensitivity of the approach to milk and meat allocation decisions.
Acknowledgements
The authors express their gratitude for financial support provided under the National Development Plan, through the Research Stimulus Fund administered by the Department of Agriculture, Fisheries and Food, Ireland, No. RSF 07 517.
