texte-dolle-anglais

INRA Prod. Anim., 2011, 24 (5), 415-432

Greenhouse gases in cattle breeding: evaluation and mitigation strategies

J.-B. DOLLE1, J. AGABRIEL2, J.-L. PEYRAUD3,4, P. FAVERDIN3,4, V. MANNEVILLE5, C. RAISON6, A. GAC6 et A. LE GALL6 1 Institut de l’Elevage, 56 Avenue Roger Salengro, BP 80039, F-62051 Saint-Laurent-Blangy, France 2 INRA UR 1213 Herbivores, F-63122 Saint-Genes-Champanelle, France 3 INRA UMR1080 Production du Lait, Domaine de la Prise, F-35590 Saint-Gilles, France 4 Agrocampus Ouest, Production du Lait, F-35590 Saint-Gilles, France 5 Institut de l’Elevage, 9 Allée Pierre de Fermat, F-63170 Aubière, France 6 Institut de l’Elevage, Monvoisin, BP 85225, F-35652 Le-Rheu, France Email: jean-baptiste.dolle@idele.fr

Livestock rearing accounts for almost 60% of greenhouse gases in the agricultural sector. This article presents the methods used to assess the carbon footprint of dairy and beef production, and analyses the main variation factors in order to identify mitigation strategies and progress margins.

In 2006, the FAO report "Livestock’s long shadow: environmental issues and options" estimated that livestock production was responsible for 18% of global greenhouse gas emissions (GHG) (FAO 2006). More recent studies have revised the values of some sectors like dairy farming. In the context of these new assessments of global GHG emissions conducted in 2010, the dairy cattle sector is estimated to account for 2.7% (FAO 2010). These assessments conducted by FAO concern the entire production chain, including feed production, herd management, manure management, milk transportation, etc.

In France, agricultural activity represented 18.8% of national GHG emissions in 2009 (excluding fossil CO2 integrated into the transport sector), of which 10% was directly linked to cattle farms, or almost 60% of emissions in agriculture, considering the areas devoted to livestock rearing (CITEPA 2011).

At European level, the Joint Research Center (JRC, Leip et al 2010) conducted a study based on life cycle assessment up to the farm gate. This study highlights that the grass-based and grain-based livestock rearing sector is responsible for 9.1% of GHG emissions, without considering land use changes (deforestation, grassland tillage) and 12.8% including land use changes. Dairy and beef livestock each contribute 29% of emissions, pig production 25% and other productions 17% (Leip et al 2010).

The processes which lead to GHG emissions in livestock rearing are com-

plex and interrelated. All carbon dioxide (CO2) and water emissions related to plant and animal respiration are considered as "biogenic" and do not constitute an additional contribution to the greenhouse effect. On the other hand, nitrous oxide (N2O) and methane (CH4), like carbon dioxide (CO2) from fossil fuel consumption, are "anthropic" and as such must be managed in the same way as emissions from other human activities.

To reduce the effect of anthropic activities on climate change, the Kyoto Protocol, ratified in 2005, aims at reducing GHG emissions at international level by 5.2% by 2012 compared with 1990, i.e. 8% for the European Union and stabilization for France. In the framework of a second commitment post 2012, the reduction objectives could be increased to 20% or even 30% by 2020 compared with 1990. In the context of the "EU Climate and Energy Package", European policy is following the lines of the Kyoto Protocol with a declared aim to reduce GHG emissions by 20% on the horizon 2020 compared with 1990, or by 30% in the event of an international agreement. To achieve these objectives, EU Member States are implementing policies aimed at integrating environmental aspects into production and consumption strategies.

Thus, in the context of the "Grenelle de l’Environnement", France plans to adopt sustainable production and consumption measures. The aim of the Grenelle 1 and 2 laws is to develop measures towards sustainable consump-

tion like reducing GHG emissions and energy consumption, and displaying the environmental impacts of mass-market products. In Germany, in a report published in 2010, the government highlights measures which can help to reduce GHG emissions. In the Netherlands, the Ministry of Agriculture has signed a commitment with leading representatives from the agricultural sector for a 30% reduction in GHG emissions between 1990 and 2020 through proactive policies. Similar action has been taken in the United Kingdom where a roadmap has been drawn up on ways to reduce GHG emissions through combined action by breeders, industry and consumers. In parallel, private initiatives to assess the carbon footprint of mass-consumer goods are being launched. The aim of these assessments is to inform administrations and consumers of the potential impact of products, particularly food products.

In ruminant production, it is useful to make an analytical assessment of emissions per gas or per activity to obtain information about emission mechanisms. However, the complexity of the processes and interaction between the different components – soil, grassland, crops, feed, animals, etc. - mean that it is also necessary to take into account the carbon sequestration mechanisms related to the presence of grassland. This problem of GHG emissions and carbon sequestration also raises the question of which type of approach to consider: an approach per sector or an overall

INRA Productions Animales, 2011, numéro 5

416 / J.-B. DOLLE et al

approach. Investigations conducted over the last few years have pointed out the need to elaborate overall assessment measures, rather than measures focused on just one gas or one practice. With an overall assessment, all the fluxes present on an agricultural holding and between the different compartments of the system (soil-animal-plant) can be considered. On this basis, it seems essential to develop overall assessments at production system level and to acquire more comprehensive knowledge of the factors which explain the variability of the environmental impacts observed and to situate the practices to be implemented on cattle farms. The methodology used is based on Life Cycle Assessment (LCA). Methodological choices must be made so that this method initially developed for industrial goods can be adapted to agricultural products. The methodological foundations of this approach are improved by research and assessment work which makes it possible to specify the GHG emissions and carbon sequestration factors, the assessment boundary, allocations between by-products, etc. Applied to contrasted production systems and to situations presenting different optimization levels, it is thus possible to identify mitigation strategies and assess their efficiency.

1 / Methodological elements implemented to assess GHG effects 1.1 / LCA-based methodology Several gases are responsible for the increase in the greenhouse effect: these are mainly carbon dioxide (CO2), methane (CH4) and nitrous oxide (N2O); there are also other gases like fluorinated gases (CFC, HFC, PFC, SF6) but these do not concern the agricultural sector. The contribution of these gases to the greenhouse effect is variable and is expressed in terms of their Global Warming Potential (GWP).

The Global Warming Potential of a gas is defined as its cumulative radiative forcing effects over a specific time horizon, in this case 100 years. This value is measured in relation to CO2 and the impact of each gas on the greenhouse effect is expressed in kilograms of CO2 equivalent (25 kg CO2/kg of CH4 and 298 kg CO2/kg of N2O). GHG emissions are then expressed in connection with the primary function represented by the product. For a product/sector approach, the Functional Unit (FU) is usually the quantity of product, often expressed in weight or volume unit (kg of gross product, kg of dry matter, kg of meat, litre of milk). It can also be a unit

INRA Productions Animales, 2011, numéro 5

related to a quality criterion or a desired functionality (kg of protein, kcal of gross energy, MJ of potential energy…).

This "GHG" environmental impact assessment is carried out in the classic manner, using the Life Cycle Assessment method, as defined in ISO standard 14040. This makes it possible to observe the direct and indirect potential environmental impacts of a product throughout its life cycle. Initially implemented for manufactured goods, the application of LCA is now being developed for the agricultural sector (Rossier and Gaillard 2001). Important scientific studies have been conducted at world level (FAO 2006, FAO 2010), at European level (Leip et al 2010), at national level (Garnett 2007, Kool et al 2009) and at agricultural product level (Cederberg and Flysjö 2004, BassetMens et al 2005, Van der Werf et al 2009, Veysset et al 2010, Kristensen et al 2011,) with the aim of assessing the impact of the agricultural sector and grass-based cattle breeding.

The implementation of LCA requires recourse to emission factors which indicate the quantity of GHG emitted by a process or an activity. To do this, the Intergovernmental Panel on Climate Change (IPCC) proposes a 3-tier emission assessment system (IPCC 2006): - Tier 1, which uses default non-country specific data and equations; - Tier 2, which applies coefficients based on country-specific or regionspecific data (more detailed activity data, country-specific coefficients); - Tier 3, which uses models and inventory measurement systems adapted to national conditions, (high-resolution activity data, higher spatial and temporal scales, etc.) to enable more accurate estimates to be made and to take greater account of the mitigation strategies implemented to reduce these emissions. However, these models must use data which are accessible at national level to calculate the inventories. They must also be recognized by an international body, failing which the results cannot be taken into account by the public decision-makers.

Following the LCA methodology and IPCC recommendations, the International Dairy Federation (IDF) has published a Tier 2 methodology adapted to the dairy sector (IDF 2010). Other international initiatives are underway, in particular for lamb production. In France, Agricultural Technical Institutes have developed the GES’TIM method (Tier 3) which proposes methods of estimation and emission factors specific to the French territory: soil and climate conditions, animal husbandry practices

and crop production techniques, French energy mix, origin and production itinerary of inputs adapted to the supply of French agricultural holdings (Gac et al 2010a).

1.2 / Analysis boundary and emission sources The application of LCA methodology to agricultural holdings involves considering all the impacts on the part of a product’s life cycle up to the farm gate.

This assessment is the first link in the overall assessment of the product’s GHG, including the upstream agricultural element, transport, processing, distribution, etc. Over the boundary up to the farm gate, an inventory should be made of the direct impacts related to the on-farm production process together with the indirect impacts inherent to the manufacture of inputs and their transport (figure 1). This usually involves productive and replacement animals, area intended for the unit and all the inputs (energy, fertilisers, feed…) used for this unit and this area (Gac et al 2010c). For milk and meat for consumption, the emissions associated with the farm boundary represent 70 to 90% of the total emissions of the full life cycle of both products (Tomasula and Nutter 2011). It is therefore essential to have thorough knowledge of the onfarm impacts in order to assess the impact on the complete life cycle of products intended for human food consumption.

The emission sources related to this boundary are listed under five main headings: - Enteric fermentation: methane emissions from animal biological activity in the unit; - Manure management: methane and nitrous oxide emissions from manure management (grazing, buildings, storage); - Nitrogen intake: nitrous oxide emissions related to nitrification and denitrification of direct nitrogen intake through organic (including crop residue) and mineral fertilization, and indirect intake resulting from nitrogen enrichment through nitrate leaching and ammonia volatilization; - Direct energy: CO2 emissions from on-farm fossil energy consumption and CO2 (electricity and fuel oil); - Inputs: impact in CO2 equivalent generated during the manufacture and transport of inputs (fertilizer, animal feed, seeds). In particular, this involves CO2 of energy origin, but also other GHG, including N2O from fertilizer manufacture.

Greenhouse gases in cattle breeding: evaluation and mitigation strategies / 417

Figure 1. Emissions and fluxes retained to assess the GHG impact.

The factors of the main emission headings used to assess French grassbased livestock rearing systems come from the GES’TIM method (table 1).

The buildings and equipment, representing less than 5% of the GHG impact, can be retained in the scope of the study (IDF 2010), but are often excluded because of the very small differences observed between the constructive techniques (Erzinger et al 2003) and the low GHG impact in relation to total emissions (Blanchin et al 2010). Emissions associated with pesticides, detergents and veterinary products which have a GHG impact of less than 1% are usually excluded from the boundary (Henriksson et al 2011).

1.3 / Land use change and carbon sequestration Beyond the GHG emissions linked to production processes, carbon fluxes whether or not related to land use change, must also be considered in the GHG impact assessment. By land use change we refer to non-agricultural land converted into agricultural land or when there has been an acknowledged agricultural land use change. These GHG

emissions or carbon sequestration related respectively to the storage or destorage of carbon must also be assessed for each of the entries into the product’s life cycle. This is particularly the case for imported feeds which are included in the composition of rations and are linked to land conversion, like soybean.

Thus for soybean cake, the data used by FAO 2010 (0.93 kg CO2/kg of soybean cake from Argentina partially associated with a conversion of grassland to cropland and 7.69 kg CO2/kg for soybean cake from Brazil and entirely associated with deforestation) or data compiled by Da Silva et al (2010) (0.694 kg CO2/kg of soybean cake from the centre-east of Brazil and 0.337 kg CO2/kg of soybean cake from southern Brazil), are very different. These differences linked to the assessment boundaries and impact allocations highlight the importance of taking these fluxes into account and the need to specify the values. Beyond imported products, carbon sequestration also involves national area subject to land use change. The results of a collective expertise conducted by INRA (Arrouays et al 2002) have revealed the average rates of storage/destorage

observed in France (table 2) following land conversion. The case of grasslands is of particular importance given their place in French cattle rearing systems.

Tillage or implantation of grasslands affects carbon fluxes. So, the conversion of forest to grassland or cropland, or of grassland to cropland, leads to CO2 emissions of up to 4.6 tonnes CO2/ha/year and of nitrous oxide due to organic matter decomposition in the soil. Inversely, the conversion of cultivated land to grassland results in carbon sequestration in the soil of 0.84 to 2.75 tonnes CO2/ha/year (Arrouays et al 2002).

Furthermore, in "stabilised" situations with no significant land use changes, some authors (Watson et al 2002) specify that the soil’s organic matter content reaches a balance, including for soils on grassland where the annual flux is nil.

Nevertheless, several publications (Bellamy et al 2005, Soussana and Lüscher 2007, Smith et al 2007) highlight the presence of an annual carbon flux on grassland areas over the long term. This would indicate that there is no temporal limit to carbon sequestra-

INRA Productions Animales, 2011, numéro 5

418 / J.-B. DOLLE et al

Table 1. Main headings and emission factors - GES'TIM method (Gac et al 2010).

LU: Livestock Unit.

tion, with very long standing grasslands continuing to store carbon over very long periods. Recent studies on carbon sequestration on grassland conducted in the context of the European Green Grass (Soussana et al 2007) and CarboEurope projects (Schulze et al 2009) show that the gross primary productivity of grassland is comparable to that of forests in Europe. Grasslands are net sinks for atmospheric CO2, storing

between 500 and 1200 kg C/ha/year depending on the management measures (animal load, method of use, fertilization). In the United States, Pelletier et al 2010 propose values of 120 to 400 kg C/ha/year on grassland. However, this soil carbon sequestration can present a certain vulnerability to climate change (Ciais et al 2003), in particular due to an increase in temperature which favours the mineralization of stored car-

bon and thereby CO2 recycling.

Nevertheless, this potential destorage phenomenon can be offset by the increase in atmospheric CO2 which reduces the grassland ecosystems’ sensitivity to drought (Morgan et al 2004) and by the productivity of grassland which increases from 5% to 15% depending on water and nutrient availability (Soussana and Hartwig 1996, Tubiello et al 2007).

Table 2. Impacts related to land use change on soil carbon storage (Arrouays et al 2002).

INRA Productions Animales, 2011, numĂŠro 5

Greenhouse gases in cattle breeding: evaluation and mitigation strategies / 419

Table 3. Proposed values of annual net fluxes of additional soil carbon sequestration, depending on land use; per hectare, on a 0-30 cm horizon, 20-year scenario (Gac et al 2010 according to Arrouays et al 2002, Soussana et al 2010).

To assess carbon sequestration, the IPCC (2006) methodology recommends considering the losses and gains in carbon stock observed during the first 20 years following land use change.

Because of a lack of data at world level, the assessments and methodologies developed to determine GHG (IDF 2010, FAO 2010) do not consider carbon sequestration in grassland systems of more than 20 years. At European level, the Joint Research Centre retains sequestration rates of 237 kg C/ha for permanent grassland and 115 kg C/ha for temporary grassland (Leip et al 2010). In a context where this point is discussed, average factors of storage/destorage (Gac et al 2010b) have been proposed for French grass-based production systems (table 3). These proposed levels are deliberately cautious

with regard to the latest published references mentioned above. Pending additional studies from EU-funded programmes to obtain more precise values, they constitute an average value with regard to the variability of situations (plain, mountain, age of the grassland, etc.) and the related degrees of uncertainty. Finally, account should also be taken of carbon sequestration by hedges and groves, found very frequently on liverstock farms. Their area, which can represent between 10 and 15% of a farm’s agricultural area (Gac et al 2010b) is an important factor. In view of all these elements, it is advisable to reason in terms of net balance, considering both GHG sources and their compensation through carbon sinks (Soussana et al 2010).

1.4 / Allocation Often made up of several production units, a farm markets several products, hence the need to look for a way to distribute the impacts amongst these different products. Given the complexity of allocation, recourse to the latter must be avoided whenever the analysis of a production system permits (ISO 14044), i.e.by dividing production processes into sub-processes and by separate data collection. This is why the analysis of an agricultural activity must take care to ventilate the fluxes of matter and emissions per unit. Two units making up a farm (cattle and crops or dairy and beef cattle‌) or two components of a unit (productive and replacement animals) must therefore be analysed separately (figure 2). This situation is found in dairy units which produce both milk and meat and which have productive animals (dairy cows, fattening animals) and animals intended for herd replacement. The problem is identical for specialized meat production systems where there is simultaneous production of culled females, grazers, young bulls, etc. Wherever possible, a distinction should be made between the production stages by allocating the emissions to the appropriate animal category. This allocation of emissions according to the production stages is to enable a distinction to be made between the different

Figure 2. Analysis level of GHG emissions per farm and per production stage.

INRA Productions Animales, 2011, numĂŠro 5

420 / J.-B. DOLLE et al

Table 4. Average ratios of partitioning of GHG impact to milk and meat according to different allocation techniques (Dollé et al not published).

* RICA (Agricultural Accounting Information Network) data (averages for 2003-2009). ** 1 kg of meat = 9.5 kg of CO2/kg live weight.

processes (figure 2). In dairy production, this implies allocating the emissions related to the growth and fattening of animals to “meat production”, and emissions related to milk cows to “milk and calf production”.

Nevertheless, when a stage in the life cycle results in several co-products (e.g. meat from culled dairy cows) and when the partitioning of emissions between the different animal categories is impossible, recourse to allocation is necessary to share the GHG impact between the milk and meat produced within these dairy units (Cederberg and Stadig 2003, de Vries and De Boer 2010, Kristensen et al 2011). Five main methods of allocation between milk and meat can be identified for dairy farms: 1/ Allocation to the main product, where the whole impact of the dairy unit is allocated to milk production; 2/ Protein allocation, on the basis of the protein content of milk and meat, this allocation is recommended by FAO 2010; 3/ Biological allocation, on the basis of the feed energy required to produce the milk and meat, this allocation is recommended by IDF 2010; 4/ Allocation by system expansion, where emissions linked to beef production (Cederberg et al 2003) or to mixed beef and pork (Nielsen et al 2003) are deducted from the emissions from the milk unit on the basis of an average carbon footprint allocated to meat; 5/ Economic allocation, on the basis of the economic income guaranteed respectively by milk and meat.

It should be noted that these emission allocation methods answer three distinct logics. The first is based on the functioning of the production system (allocation per production stage and biological allocation), the second relies on keys linked to the nutritional value of the final product (protein allocation) and the third is based on elements of context (allocation to the main product, allocation by system expansion, economic

INRA Productions Animales, 2011, numéro 5

allocation). The choice between these keys for allocation of GSG must be made in order to be legible by the different players in the impact calculation but also for the implementation of action plans. The techniques used for allocation based on the functioning of the exploitation system (allocation per production stage and biological allocation), are thus of interest.

Depending on the allocation technique chosen and the boundary on which it is applied, the allocation of GHG emissions between milk and meat is different (Dollé et al not published).

For French dairy units producing milk, calves and culled females, the GHG impact allocated to milk is between 72% and 100% (table 4). This impact distribution between milk and meat varies according to the economic context, the milk/meat ratio (Kristensen et al 2011), zootechnical performances and a fortiori for protein and biological allocation techniques when the analysis boundary is different. The problem of allocation which focuses here mainly on the dairy herd can in some cases involve the suckler herds producing several types of meat (culled animals, oxen, young bulls, heifers) Dollé et al (2011).

The principles presented above (partitioning per production stage or allocation) apply to these systems in an identical way.

2 / Overall assessment of the livestock system Livestock rearing farms are complex systems and there is multiple interactions between the herd and the areas mobilized. An overall assessment must be made at system level to obtain a representative picture of the environmental impact and to identify mitigation strategies to reduce this impact. This overall assessment of GHG emissions per kilogram of product enables the gross carbon footprint to be calculated. By taking account of carbon sequestration, the net carbon footprint can be assessed.

2.1 / In dairy production systems Extensive studies have been carried out in France to assess the carbon footprint of the milk produced in leading dairy production systems, following the GES’TIM methodology described above. On a sample of 153 conventional French dairy farms belonging to three different production systems (figure 3), the gross average carbon footprint observed is 1.27 kg CO2/l of milk after application of the protein allocation (Dollé et al not published). The high methane contribution in the milk carbon footprint (63%) is linked to the preponderant share of enteric fermentation (69% of CH4 emissions). The other emissions are divided between nitrous oxide (17%), influenced by emissions during grazing, representing 41%, and carbon dioxide (20%), produced by onfarm fossil energy combustion and the impact of inputs. These emission levels are coherent with the bibliography which reveals gross carbon footprints ranging from 0.8 and 1.5 kg CO2 per litre (table 5). Some differences are due to the systems studied, to specific cases with few farms, to the methodological choices (boundary, allocation rule, calculation method) and to the functional unit (gross milk or fat and protein corrected milk). The results are nevertheless fairly similar considering the wide diversity of the world systems studied, in particular if one considers the differences between the intensive grasslandbased grazing systems in New Zealand and Ireland and the stock-based systems in the Netherlands.

The analysis of the results obtained from the three French production systems (figure 3) indicates very slight variations in gross CO2 emissions from one system to another. The lower carbon impact from higher productivity systems on plains (7 718 kg milk/cow/year) is not significantly different from the results obtained from the other production systems. It is important to note that higher productivity per cow, resulting in lower enteric methane impact per litre of milk, does not sys-

Greenhouse gases in cattle breeding: evaluation and mitigation strategies / 421

Figure 3. Gross and net footprints for milk from 3 French dairy systems (Dollé et al not published).

tematically lead to a lower carbon footprint at system level (Martin et al 2011).

The increase in productivity on enteric methane expressed per litre of milk usu-

ally implies a higher use of inputs (concentrates, fertiliser), higher direct energy (fuel oil) consumption related to forage stocks responsible for higher

carbon dioxide emissions and higher emissions in the buildings. Therefore it is important to note that by only assessing enteric methane emissions, it is not possible to reach a conclusion on the system’s overall emissions. This absence of any significant effect of productivity or the intensity level on the carbon footprint per kg of milk also explains the small differences observed between conventional and biological systems (Cederberg et al 2000, Haas et al 2001, Van der Werf et al 2009, Kristensen et al 2011, Chambaut et al 2011). So a higher methane emission (expressed in kg CO2/kg of milk) due to lower productivity by the biological systems is usually offset by lower carbon dioxide and nitrous oxide emissions because of less recourse to inputs. The effect of the productivity level on the gross carbon footprint expressed per litre of milk is nevertheless observed for systems with productivity/cow of less than 4000-5000 kg of milk (Gerber et al 2010) (figure 4), to the contrary of systems where productivity/cow is over 5000 kg of milk (Doreau and Dollé 2011a, Vellinga et al 2011). Furthermore, grassland area which guarantees soil carbon sequestration represents variable compensation of GHG emissions of between 6 and 43% with a sequestration hypothesis of 500 kg C/ha/year (figure 3). The highest carbon sequestration rate obtained in systems with over 95% grassland totally offsets methane of enteric origin (figure 3).

Table 5. Comparison of the carbon footprint for milk production according to different bibliographical sources.

Spec.: specialized; Conv.: conventional. BA: Biological Agriculture.

(1): uncorrected milk; (2): fat and protein corrected milk; (3): energy corrected milk.

INRA Productions Animales, 2011, numéro 5

422 / J.-B. DOLLE et al

Figure 4. Relation between GHG emissions and the productivity of dairy cows (Gerber et al 2011).

Within the typological category “specialized milk from plains”, a classification of French dairy farms has been established on the basis of net carbon footprint results (table 6).

Within this population of 127 farms, a 28% difference in the net carbon footprint of milk is observed between the upper and lower quarters and a 12% difference on the carbon footprint of milk between the upper quarter and the intermediate class. As observed by Henrikson et al 2011, the variations in the carbon footprint are strongly related to herd management and crop production practices. This explains why the most optimized systems which have the lowest consumption of concentrates per litre of milk but also the most efficient nitrogen balances are the most favourable environmentfriendly. Finally, the economic and environmental results of this analysis bring to light the relation between economic and environmental efficiency.

The failure of technical optimization which leads to a higher carbon footprint generates lower economic results (table 6).

2.2 / In meat production systems Studies carried out in France to assess meat production from conventional suckler herds (Dollé et al 2009, Gac et al 2010c, Dollé et al 2011, Veysset et al 2011) show a gross carbon footprint of between 9.5 and 17.8 kg CO2/kg live weight (table 7), with no distinction as to the type of meat produced (culled cows, young bulls, oxen…). The contribution of the three main GHG highlights the preponderant share of methane (71%). Enteric fermentation represents 52% of total emissions, or 73% of total emitted methane.

Emissions in the storage buildings account for 21%. The emissions of N2O represent 17% of the total footprint and are mainly related to grazing (42%).

Emissions of CO2 due to the consumption of direct energy and chiefly to inputs, account for 12%. These gross carbon footprints in French bovine livestock production assessed using two different methodologies, GES’TIM for Gac et al (2010a) and Dollé et al (2011); and PLANETE-OPT’INRA for Veysset et al (2011), are lower than the assess-

ment for beef at European level, which is 22.2 kg de CO2/kg live weight (Leip et al 2010). However, this average dissimulates significant differences.

Austria (14.2 kg CO2/kg live weight) and the Netherlands (17.4 kg CO2/kg live weight) have the lowest gross carbon footprint compared with Cyprus (44.1 kg CO2/kg live weight) and Latvia (41.8 kg CO2/kg live weight) where the lower efficiency of production systems is linked to a significant land use change. Thus balanced systems coupled with optimized fattening periods (Benchaar et al 2001, Lovett et al 2005, Casey and Holden 2006, Dollé et al 2011) enable a reduction in the carbon impact per kg of meat production. The effect of productivity is also highlighted

by Pelletier et al (2010), where the gross carbon footprint of beef cattle fattened in the USA on non-optimized grazing systems is higher than beef cattle fattened in feedlots (19.2 kg CO2/kg live weight against 14.8 kg CO2/kg live weight) due to a lower daily weight gain (0.6 kg/day against 1.4 kg/day).

This carbon footprint of meat from suckler herds includes the rearing and fattening stages of grazers. The highest share of emissions from these production systems is from suckler cows. The breeder systems, where cows account for the main part of the animal stock and where the quantity of valorized meat is lowest, thus have a higher carbon footprint compared with breeding-fattening farms. At the fattening stage alone, Doreau et al (2011b) have highlighted a carbon footprint of between 3.65 and 4.74 kg CO2/kg in live weight gain, depending on the share of forage and concentrates in the ration; compared with breeder-fattener systems which show a gross carbon footprint of between 12.8 and 17.8 kg CO2/kg live weight (Gac et al 2010c, Veysset et al 2011, Dollé et al 2011). This is why GHG emissions linked to the rearing of grazers represent 75% (Dollé et al 2009) to 80% (Phetteplace et al 2001) of the final footprint. The suckling stage together with the mother’s conduct is the main GHG source over the animal’s entire life cycle and therefore has a con-

Table 6. Environmental and economic efficiency of 64 farms belonging to specialized dairy systems with more than 10% maize/ha Principal Forage System (Raison et al not published).

INRA Productions Animales, 2011, numéro 5

Greenhouse gases in cattle breeding: evaluation and mitigation strategies / 423

Table 7. Comparison of the carbon footprint from meat production according to different bibliographical sources.

(1) Meat from young dairy cow.

siderable influence on the carbon footprint of the final product.

15% lower than intensive systems (Pelletier et al 2010).

Few studies have been conducted so far to determine the net carbon footprint taking account of carbon sequestration for meat systems. Work by Dollé et al (2009), Dollé et al (2011), Gac et al (2010b) shows that between 24% and 53% of GHG emissions are offset by carbon sequestration from grassland fixed at 500 kg C/ha/year. On the basis of carbon sequestration by grassland of 350 kg/ha/year, Veysset et al (2011) show a compensation of between 13.3% and 21.2% depending on the fraction of grassland valorized in the system. This is how more intensive systems (high share of concentrates, low grazing systems, high daily average weight gain), which have a lower gross carbon footprint expressed in kg CO2/kg live weight, are associated with low carbon compensation and thereby result in a net carbon footprint which is identical or even higher than for other systems (Veysset et al 2011, Dollé et al 2011).

Pelletier et al (2010) who have used the values proposed by Phetteplace et al (2001) indicate carbon compensation of 42% for grazing systems against 12% for beef cattle fattened in feedlots.

Optimized pasture systems would thus appear to have a net carbon footprint

The conversion of the production system to biological agriculture (Veysset et al 2011, Chambaut et al 2011) leads to a lower net carbon footprint, in particular due to the importance of grassland in the biological system. Without taking carbon sequestration into account, Casey et Holden (2006) indicate an emission rate expressed per kg of live meat which is 14% lower for biological systems whereas Haas et al (2001) do not note any significant difference between the two production methods.

2.3 / Meat produced in dairy systems Although milk is the principal product from dairy production systems, meat production nevertheless represents a significant share estimated at 40% of the beef produced in France. On a dairy unit, where replacement animals are raised jointly with productive animals, the emissions must be partitioned between the two products: milk and meat. Many of the above-mentioned methods to allocate GHG emissions have been applied to dairy systems (Dollé et al not published) to determine the net carbon footprint for milk and

meat respectively from culled animals.

The effect of the different allocation techniques on the carbon footprint is very strong (figure 5). The partitioning of emissions per production stage, which involves allocating the emissions from replacement animals to “meat” and the emissions from the lactation and dry stages to “milk and calves”, leads to a carbon footprint for milk of 0.83 kg CO2/kg and a carbon footprint for meat from culled dairy cows of 8.7 kg CO2/kg live weight. Allocating GHG emissions according to the animals’ food requirements (IDF, 2010) for milk and meat production (biological allocation), gives a carbon footprint for milk of 0.9 kg CO2/kg and for meat of 6.4 kg CO2/kg live weight. These two methods of sharing (biological allocation and partitioning per production stage) in relation to the production system, lead to a carbon footprint which is closer to the system expansion allocation where the footprint of meat corresponds to the average for meat from specialized dairy systems. Application of the protein allocation as recommended by FAO (2010), gives a lower carbon footprint for meat (4.4 kg CO2/kg live weight). Finally, allocation to the main product, where all the emissions are allocated to the produced milk gives the highest carbon footprint for milk (1.1 kg CO2/kg milk).

Apart from the results obtained using

INRA Productions Animales, 2011, numéro 5

424 / J.-B. DOLLE et al

Figure 5. Net carbon footprints for milk and meat from dairy farms according to different allocation techniques (Dollé et al not publiished).

the economic allocation, where the milk/meat ratio is greatly dependent on the economic situation and the income retained (with or without aids), the calculated footprints are in line with the results obtained by Kristensen et al (2011) who carried out the same assessment.

Since a significant fraction of GHG emissions produced by a dairy cow in lactation are allocated to milk, the calves from dairy herds have a lower carbon footprint than that produced in the suckler system. An assessment of French systems specialized in the fattening of young bulls from dairy herds shows a gross carbon footprint of 4.5 kg CO2/kg live weight. These results corroborate the results obtained by Cederberg and Darelius (2002) where the carbon footprint is 6.45 kg CO2/kg live weight. It should be noted that in all cases, the specialized milk systems and mixed milk/meat systems give a lower net carbon footprint for meat than that from specialized meat systems.

3 / Mitigation strategies To meet national, European and international objectives, technical recommendations are proposed (Metz et al 2007). The differences observed between production systems and farms from the same system (Henriksson et al 2011) bring to light potential strategies to reduce the carbon footprint of milk and beef products. Strategies exist to reduce methane, nitrous oxide and carbon dioxide emissions. These strategies must be assessed and tested at system level and for the three GHG concerned in order to avoid any pollution transfer from one gas to another or from one compartment to another.

INRA Productions Animales, 2011, numéro 5

3.1 / Feed Animal feed is a mitigation strategy to reduce enteric methane emissions and nitrogen discharges from animals.

An increase in the portion of concentrates in the ration, which replace the carbohydrates from the forage plant cell walls with starch and sugars, is frequently quoted as a means of reducing enteric methane emissions (Martin et al 2009, Doreau et al 2011a). However, Hindrichsen et al (2006) indicate that emissions from the excreta of dairy cows fed with a diet of forage and feed concentrate are higher than for a forageonly diet. Apart from the effect on methane, this technical option, which consists of increasing the concentrate portion of the ration, can result in higher production costs and put ruminants in competition with Man for access to food resources. Furthermore, concentrate production itself also has a GHG impact which makes this solution less attractive than a rapid analysis may lead to believe. Thus on French dairy farms (table 6), there has been 24% over-consumption of concentrates between the non-optimized farms with the highest carbon footprint and the optimized farms with the lowest carbon footprint (Raison et al not published). Over the fattening stage alone, this increase in enteric methane from high concentrate content diets (> 70%) remains established at system level, despite higher N2O and CO2 emissions (Doreau et al 2011b). Another strategy to reduce enteric methane emissions is by ration enrichment with unsaturated lipids, which enables an average 3.8% reduction in methane emissions for a 1% lipid addition to the ration (Martin et al 2009). The effect on enteric methane is nevertheless highly dependent on the type of lipid used (lauric, myristic, linoleic acids) (Martin et al 2009).

Some authors (Woodward et al 2006) question the long term effect of lipids

on methane emissions but recent data obtained with linseed clearly show a positive long term effect on enteric emissions (Martin et al 2011). Food additives (ionophores, organic acids) could be another way to reduce enteric methane emissions, but their development is still subject to numerous queries (Doreau et al 2011a). Some substances are prohibited in Europe (ionophorous antibiotics), others are of chemical origin and very costly (organic acids), and others have not proved their lasting effectiveness (tannins, saponins).

With regard to animal nitrogen discharge, an assessment has revealed that a reduction in the nitrogen content of the diet seems less effective than action on fertilizer management (Henriksson et al 2011). However, a reduction in fecal and particularly urinary losses (Kebreab et al 2001) is an interesting path to limit N2O losses related to animal waste management and also CO2 losses related to input production.

Therefore, finding the correct balance between contributions and requirements provides a very interesting and simple path to reduce the consumption of concentrates and nitrogen losses. In some cases, it even seems possible to reduce the applications of degradable nitrogen to slightly below the recommended levels to increase the efficiency of nitrogen use and further reduce losses in the form of urea without penalizing performances too much. This manoeuvring margin can be used for ruminants well-fed with amino acids (Vérité et al 1997). A 5% shortage in degradable nitrogen intake compared with recommended levels will not have any consequences on production efficiency but will enable a reduction of about 10 kg of N discharge per lactating cow. A greater significant reduction will lead to a significant decrease in efficiency.

Another possible path to follow regarding animal feed is to buy feed with a low carbon impact in relation to growing practices or land use change (rapeseed versus soybean). Finally, if a farm can achieve feed autonomy, this can reduce GHG emissions. This particularly involves protein autonomy on farms by using grass and leguminous crops. Each of the feed strategies mentioned here will allow a reduction of 0 to 8% in GHG emissions, depending on the case (table 8). Extensive research is nevertheless still necessary to determine their environmental and techno-economic interest.

3.2 Productivity and herd management An increase in dairy productivity is frequently mentioned as a way to reduce

Greenhouse gases in cattle breeding: evaluation and mitigation strategies / 425

Table 8. Synthesis of mitigation strategies and potential gain on GHG emissions.

the enteric methane impact per kg of product and could allow a reduction in carbon footprint from 0.03 kg CO2/kg milk (Schils et al 2005) to 0.26 kg CO2/kg milk (Vellinga et al 2011).

However, this trend is questioned depending on the production system studied (Gerber et al 2011) and the productivity level taking account of the carbon weight of inputs, higher gas emissions in the cattle rearing buildings as opposed to the grazing period, and lower carbon compensation from grassland. The carbon footprint assessment made on a sample of 153 French dairy farms (DollĂŠ et al not published) shows a slight trend towards a reduction in the gross carbon footprint with the increase in productivity (figure 6a). On the same sample, including carbon sequestration in the net footprint calculation highlights the negative impact that an increase in productivity could have to

the detriment of area under grassland (figure 6b).

Furthermore, Lovett et al (2006) specify that the increase obtained on enteric methane from high productivity cows can be lost if we consider the cow’s entire life and the number of replacement females required. Thereby, rearing situations with a generally lower number of lactations and replacements and animals which are more sensitive to health risks, are associated with a higher replacement rate and consequently with a higher percentage of non-milk producing animals. Since the share of GHG emissions from a replacement herd represents 25 to 35% of dairy unit emissions, any action to reduce the proportion of animals grown and maintained as opposed to productive animals must be studied. A reduction of 5 to 9% in the replacement rate could enable a

reduction in the on-farm carbon footprint of 20 to 39 g CO2/kg of milk (Vellinga et al 2011). However, these observations are only relative since the gains noticed on dairy farms depend on the method of partitioning or allocation of the impact between milk and meat.

Consequently, any action taken on dairy herd replacement can have an effect on the carbon footprint of milk with regard to protein, economic allocation, etc. but will have no effect on the final impact of milk in the case of partitioning per production stage or biological allocation. This means that at overall level for livestock farms producing milk and meat, these strategies to reduce the carbon footprint for milk by increasing productivity or reducing the replacement rate on the dairy unit would have no effect on total GHG emissions. In order to meet the same demand for milk and meat, any efforts focusing on an

INRA Productions Animales, 2011, numĂŠro 5

426 / J.-B. DOLLÉ et al

increase in dairy yield and the number of lactations would go against the “coproduced meat” of milk and require higher meat production from the specialized units with a higher carbon footprint (Zehetmeier et al 2011). The situation is different for beef cattle at the fattening stage. As long as this yield increase does not imply high use of inputs, it can be of interest insofar as a higher daily average weight gain allows a significant reduction in the fattening period and thereby in the associated emissions.

Figure 6. Gross (6a) and net (6b) carbon footprints for milk according to productivity per cow on 153 dairy farms (Dollé et al not published).

Finally, better herd management (culling the less productive animals, efficient health management, ability to adapt to environmental change…) represents a potential to reduce GHG emissions in livestock by between 2 and 5% (Cruickshank et al 2009). Other practices which focus more on management of the suckler herd, such as the date for calving according to the date of maximum grass growth to optimize grazing, still need to be assessed.

Genetic improvement also provides an opportunity to reduce GHG emissions. By basing breeding on improved feeding efficiency, Alford et al (2006) estimate the potential reduction in methane emissions at 3.1% over 25 years. Similarly, production efficiency obtained by better fertility, optimum ration valorizations, a reduction in maintenance costs and improved animal health (Hegarty 2004) is an important path for progress. Based on data from the bibliography and on first assessments made on French systems, all actions taken on productivity and herd management can, depending on the case, lead to a reduction in the carbon footprint of 0 to 10% (table 8).

3.3 / Nitrogen fertilization De Klein and Eckard (2008) suggest that appropriate nitrogen management can reduce nitrous oxide emissions by 90%. Although this value may seem an overestimate, it can be confirmed that nitrous oxide emissions can be greatly reduced and a wide variety of solutions explored (Luo et al, 2010). The main indicator in nitrogen management is onfarm surplus nitrogen, the difference between nitrogen intake from inputs and the nitrogen mobilized in the system products. This surplus represents a risk of leaching, nitrogen losses through ammonia volatilization and through nitrous oxide emissions. Regarding fertilization, good management and optimum agronomic valorization of animal waste can allow a considerable reduction in mineral fertilizer purchases. The in/out balance of nitrogen shows that 80% of French grass-based farms have a

INRA Productions Animales, 2011, numéro 5

nitrogen surplus of less than 90 kg/ha (Le Gall et al 2009) in line with low nitrate losses. Considerable efforts have been made over recent years (manure storage, fertilization plans…) to reduce organic mineral use; but these efforts must be continued. As well as reducing N2O emissions, any optimization of nitrogen fertilizer reduces the need to buy synthetic fertilizers and thereby reduces CO2 emissions related to their manufacture and transport, i.e.between 5.3 and 6.1 kg CO2/kg N. The use of leguminous crops is also a way to reduce N2O emissions through symbiotic fixing and a reduction linked to aerobic conditions. IPCC considers that symbiotic fixing does not emit GHG (IPCC, 2006). Corré et al (2002) highlight N2O emissions measured on grassland with lower associations than those measured on graminae grassland (0.2 against 1.3% N). Ledgard et al (2009) show that milk produced on

grassland fertilized with 200 kg of nitrogen gives a carbon footprint 15% higher than milk produced on non-fertilized grassland growing leguminous feed crops. The latter are therefore an efficient strategy to reduce the use of synthetic fertilizer. The main difficulties concern the perennity of white clover in grassland and the over-seeding of white clover on installed grassland.

Nitrification inhibitors, which aim at slowing down the nitrate production from the conversion of ammonium nitrate from animal urine and excreta, can also reduce losses through leaching and N2O emissions. These products are not very widely used in France but are studied extensively in New Zealand (Monagham et al 2007) where they are spread on grassland or mixed with organic products before application.

Some studies show a reduction in GHG emissions of 50 to 68% on urine during grazing thanks to the application of

Greenhouse gases in cattle breeding: evaluation and mitigation strategies / 427

nitrification inhibitors (Di and Cameron 2010, Monaghan et al 2007 and 2009), i.e. up to 5% of the carbon footprint of the final product. Other studies give much less positive results or sometimes indicate no effect by nitrification inhibitors (Mc Donald 2010). Since the results vary according to the experimental sites - the application conditions (date and dosage) and supports - work will be carried out to determine the efficiency of these nitrification inhibitors in France. The three possibilities mentioned here can lead to a reduction in GHG emissions and the carbon footprint of 0 to 5% (table 8).

3.4 / Waste management Animal manure management includes excreta produced during grazing and excreta produced in the farm buildings. Emissions from excreta on grazing are significantly lower than from excreta produced in the farm buildings which are then stored (IPCC, 2006). A leading path in effluent management can therefore be to reduce the quantities in storage and to increase grazing. Considerable investment has been made over recent years to optimize on-farm manure management (collection, long term storage management) and its agronomic valorization.

This improved management involves making applications during the appropriate periods, corresponding as much as possible to the plant’s needs, so that mineral nitrogen applications can be reduced. On-farm manure management can be completed by anaerobic digestion of animal waste to avoid methane emissions during storage.

Biogas, a mixture of CH4 and CO2, is valorized in energy form. In dairy systems, the installation of a methanisation unit can enable a reduction of the carbon footprint of between 5 and 7% (Dollé et al not published). But this type of unit which is subject to a number of parameters like constant manure production, availability of co-substrates, heat valorization, can only concern a limited number of installations.

The type of waste produced - slurry or manure - also affects methane and nitrous oxide emissions in buildings and on storage. Conditions for manure production and storage (on platforms or under the animals) have an effect on the level of emissions due to the degree of anaerobiosis. Since there are few available bibliographical references on waste management in the form of manure, studies are underway on the different storage facilities found in France (scraped area and straw bedding) to specify mechanisms and emission levels.

3.5 / Reducing energy consumption Emissions of "energy" carbon dioxide comes from fuel and electricity consumption. Very interesting technical solutions to reduce electricity consumption (milk pre-cooling, heat recovery, thermal solar) in fact only give a gain in carbon footprint of less than 1% due to the nuclear origin of electricity produced in France (Dollé et al not published). Action to reduce fuel oil consumption can include the development of production systems and practices.

Grazing provides one of the paths towards a reduction in fuel oil consumption used for crops, harvesting, forage distribution and spreading. Changes in behaviour and practices (economical driving, tractor settings, reduced transport, simplified growing techniques…) are also of significant interest and can enable a 1 to 2% reduction in GHG emissions.

3.6 / Carbon sequestration Carbon sequestration has considerable potential to abate GHG emissions in grassland livestock systems (Soussana et al 2010). In addition to conserving the soil carbon stock, grassland management practices can have an effect on the physicochemical conditions of the environment, the protection and increase of soil organic matter. So a moderate intake of nitrogen (fertilizer, excreta) increases carbon sequestration; on the other hand, a nitrogen shortage can lead to carbon destorage since microbes in the soil draw from the soil’s humic reserves (Fontaine et al 2004, Klumpp et al 2009). The type of plant cover also affects an agro system’s carbon sequestration capacity (Loiseau et al 2001). The presence of leguminous crops allows the self-regulation of nitrogen and helps to maintain the carbon stock (Soussana et al 2010, Loiseau et al 2001). Grazing allows better carbon sequestration than cutting via the direct intake of organic matter from excreta and a lower export of carbon because of residue grass (Reeder and Schuman 2002, Soussana et Lüscher 2007, Soussana et al 2010). Inversely, exclusive and frequent cutting of grassland (Klumpp et al 2007) can lead to too much being exported (significant and repeated export of carbon) and carbon destorage. Finally, the intensity of grazing also affects the sequestration levels because of residue grass. Grassland which is cut very short leaves fewer airborne organisms, which are sources of bedding then carbon, and over-grazing can deteriorate the plant cover (Jones and Donnelly 2004). Inversely, less severe grazing leaves a higher population of senescent organisms. (Louault et

al 2005), which enhance carbon sequestration. Finally, the duration of rotation of temporary grassland or the conversion of temporary grassland to permanent grassland increases carbon sequestration. For these different practices, a reduction in nitrogen fertiliser applications on intensive plots, conversion of graminae pastureland to mixed grass/leguminous pastureland, moderate intensification of poor grassland, the annual flux of stored carbon is between 200 and 500 kg C/ha (Soussana et al 2010). On the basis of these practices which enhance or hinder carbon sequestration, it is necessary to establish the best compromise between animal performances and carbon sequestration by grassland.

Furthermore, the conservation and planting of hedges on grassland also has significant potential for soil carbon sequestration over long periods. With regard to crops, the main variation factors are a reduction in tillage, or even a move towards no tillage which limits the carbon destorage phenomenon (the beneficial effect of a move to no tillage is nevertheless reversible). In addition, returning crop residue to the soil or spreading organic waste, in the same ways as planting catch crops to avoid fallow land (with no export of residue) increases the intake of organic matter and enhances carbon sequestration.

Overall, a change in practices or a development of area under grassland within the farms enables a reduction in the carbon footprint of between 3 and 10% depending on the system (table 8).

Conclusion Livestock rearing contributes significantly to GHG emissions. These emissions must be assessed so that action plans can be elaborated to reduce GHG impact on climate change. However, it must be underlined that ruminant rearing has the particular feature of being able to partially offset GHG emissions through carbon sequestration on grassland and agro-ecological structures (hedges and copses….). Soils are recognised as being the earth’s biggest carbon sink (1500 billion tonnes of carbon, IPCC 2007) and optimum grassland management is a promising path towards reducing GHG emissions (Arrouays et al 2002, FAO 2010, Soussana et al 2010).

The overall analysis of the livestock rearing system is crucial to make a comprehensive assessment of GHG emissions from the sector and the carbon footprint of agricultural products up to the farm gate. The methodological

INRA Productions Animales, 2011, numéro 5

428 / J.-B. DOLLE et al

strategies being elaborated should ultimately enable a harmonized quantification of these impacts. Extensive national and international exchanges are still necessary over the coming years to perfect these methodologies and the related references. The assessments carried out on French and foreign production systems show variability in the inter-system carbon footprint but even more specifically show significant variability in the optimization of practices. The differential observed can be as much as 30% between optimized and non-optimized systems. This highlights the possible progress margins for some farms by optimizing practices.

ing of the farm; in others they can sometimes require heavy investment.

Overall, for the bovine sectors, an action plan combining a range of compatible reduction strategies can provide potential to reduce GHG emissions by 5 to 15% for farms with a carbon footprint within the currently observed average.

This reduction potential emphasizes the ability of livestock rearing to face the challenge of climate change. One must also bear in mind the fact that the objectives of reaching a 20%, 30% or even greater reduction are difficult to achieve in the present state of knowledge but, in time, technological progress and genetic research should be able to provide the answers.

At the same time, a significant number of strategies which modulate emissions have been identified. The CO2 that can be avoided by applying these strategies is estimated at between 0 and 10% (table 8). Some strategies require complementary research work to assess their efficiency, reliability and cost.

Others, related to tested practices are now available and applicable to the farm. In some cases, their implementation can allow savings on the function-

It can now be specified that all the investigations conducted to assess GHG emissions and to find mitigation strategies make it possible to: - provide livestock breeders and their technical and administrative environment with the knowledge, tools and methods to guide their technical choices and modify their production systems with the aim to reduce GHG emissions and increase carbon sequestration,

- clearly indicate to livestock breeders the positive or negative economic repercussions generated by the application of on-farm mitigation strategies, - indicate to the dairy and beef sector and industries the carbon footprint that can be obtained at the farm gate, - offer France the possibility to reduce the GHG impact of its livestock rearing activity.

Finally, the environmental impact assessment of the livestock rearing activity must not be limited to the GHG aspect alone. Without bringing into question the need and urgency of implementing action plans to limit GHG emissions, a single criterion approach does not seem pertinent and can lead to pollution transfers (water pollution, atmospheric acidification). Furthermore, ruminant rearing systems offer other advantages, like contributing to the maintenance of biodiversity which must be included in the elaboration of action plans for the sectors.

Références Alford A.R., Hegarty R.S., Parnell P.F., Cacho O.J., Herd R.M., Griffith G.R., 2006.

The impact of breeding to reduce residual feed intake on enteric methane emission from the Australian beef industry. Aust. J. Exp. Agric., 46, 813-820.

Arrouays D., Balesdent J., Germon J.C., Jayet P.A., Soussana J.F., Stengel P., (Eds), 2002. Contribution à la lutte contre l’effet de serre. Stocker du carbone dans les sols agricoles de France ? Expertise scientifique collective, INRA, synthèse, 36p.

Basset-Mens C., Ledgard S., Carran A., 2005. First life cycle assessment of milk production from New Zealand dairy farm systems.

Anzsee conf., 258-265.

Basset-Mens C., Ledgard S., Boyes M., 2007. Eco-efficiency of intensification scenarios for milk production in New Zealand. Ecol.

Econom., 68, 1615-1625.

Bellamy P.H., Loveland P.J., Bradley R.I., Lark R.M., and Kirk G.J.D., 2005. Carbon losses from all soils across England and Wales 1978-2003. Nature, 437, 245-248.

Benchaar C., Pomar C., Chiquette J., 2001.

Evaluation of dietary strategies to reduce methane production in ruminants: a modeling approach. Can. J. Anim. Sci., 81, 563-574.

Blanchin J.Y., 2010. Energie et gaz à effet de serre liés à la construction des bâtiments d’élevage. Institut de l’Elevage, Paris, France, 8p.

Casey J.W., Holden N.M., 2006.

Quantification of GHG emissions from suckerbeef production in Ireland. Agric. Syst., 90, 7998.

INRA Productions Animales, 2011, numéro 5

Cederberg C., Mattsson B., 2000. Life cycle assessment of milk production-A comparison of conventional and organic farming. J. Clean.

Prod., 8, 49-60.

Cederberg C., Darelius K., 2002. Using LCA methodology to assess the potential environmental impact of intensive beef and pork production. Dep. Appl. Environ. Sci., Gotëborg University, Sweden.

Cederberg C., Stadig M., 2003. System Expansion and allocation in life cycle assessment of milk and beef production. Int. J. LCA., 350-356.

Cederberg C. and Flysjö A., 2004. Life Cycle inventory of 23 dairy farms in SouthWestern Sweden, Sik rapport, 728, 63p.

Chambaut H., Moussel E., Pavie J., Contard J.P., Galisson B., Fiorelli J.L., Leroyer J., 2011.

Profils environnementaux des exploitations d’élevage lait et viande en agriculture biologique et conventionnelle : enseignements du projet CedABio, 4p.

Ciais P., Reichstein M., Viovy N., Granier A., Ogee J., Allard V., Aubinet M., Buchmann N., Bernhofer C., Carrara A., Chevallier F., De Noblet N., Friend A.D., Friedlingstein P., Grunwald T., Heinesch B., Keronen P., Knohl A., Krinner G., Loustau D., Manca G., Matteucci G., Miglietta F., Ourcival JM., Papale D., Pilegaard K., Rambal S., Seufert G., Soussana J.F., Sanz M.J., Schulze E.D., Vesala T. and Valentini R., 2005. Europe-wide reduction in primary productivity caused by the heat and drought in 2003. Nature, 437, 529-533.

Corré W.J., 2002. Agricultural land use and emissions of methane and nitrous oxide in

Europe. Plant Res. Int., Wageningen. The Netherlands, 28p.

Cruickshank G., Thomson B., Muir P., 2009.

Effect of management change on methane output within a sheep flock. Proc. NZ Soc. Anim.

Prod., 69, 170-173.

Da Silva V.P., Van Der Werf H.M.G., Spies A., Roberto Soares S., 2010. Variability in environnemental impacts of Barazilian soybean according to crop production and transport scenarios. J. Environ. Manage., 91, 1831-1839.

De Klein C.A.M., Eckard R.J., 2008.

Targeted technologies for nitrous oxide abatement from animal agriculture. Austr. J. Exp. Agric., 48, 14-20.

De Vries M., De Boer I.J.M., 2010.

Comparing environmental impacts for livestock products : A review of cycle assessments. Livest. Sci., 120, 1-11.

Di H.J., Cameron K.C., 2010. An effective mitigation technology for reducing nitrous oxide emissions from grazed grassland. Banff, Canada. Greenhouse Gases Anim. Agric.

Conf., October 3-8.

Dollé J.B., Gac A., Le Gall A., 2009.

L’empreinte carbone du lait et de la viande bovine. Renc. Rech. Ruminants, 16, 233-236.

Dollé J.B., Manneville V., Gac A., Charpiot A., 2011. Emissions de gaz à effet de serre et consommations d’énergie des viandes bovines et ovines françaises : revue bibliographique et évaluations sur l’amont agricole, Institut de l’Elevage, Paris, France, 59 p.

Doreau M., Dollé J.B., 2011a. Strategies for reducing greenhouse gas emissions in dairy

Greenhouse gases in cattle breeding: evaluation and mitigation strategies / 429

production. A European perspective. Proc.

Eastern Nutr. Conf., Montréal, Canada, 12-13 avril 2011, 57-77.

Doreau M., Van Der Werf H.M.G., Micol D., Dubroeucq H., Agabriel J., Rochette Y., 2011b.

Enteric methane production and greenhouse gases balance of diets differing in concentrate in fattening phase of a beef production system.

J. Anim. Sci., 89, 2518-2528.

IPCC 2006. IPCC Guidelines for National Greenhouse Gas Inventories.

Intergovernmental panel on climate change.Nat. Greenhouse Gas Inv. Ref. Manual, vol. 4.

Jones M.B., Donnelly A., 2004. Carbon sequestration in temperate grassland ecosystems and the influence of management, climate and elevated CO2. New Phytol., 164, 423-439.

Erzinger S., Dux D., Zimmermann A., Badertscher Fawaz R., 2003. LCA of animal products from different housing systems in Switzerland: relevance of feedstuffs, infrastructure and energy use. In: Life Cycle Assessment in the Agri-food Sector. Halberg N. (Ed.). Danish Inst. Agric. Sci., Bygholm, Denmark, 55-63.

Kebreab E., France J., Beever D.E., Castillo A.R., 2001. Nitrogen pollution by dairy cows and its mitigation by dietary manipulation.

Nutr. Cycl. Agroecosyst., 60, 275-285.

FAO 2006. Livestock's Long Shadow: environmental issues and options. FAO, Rome, Italy, 36p.

Klumpp K., Fontaine S., Attard E., Gleixner G., Leroux X., Soussana J.F., 2009. How land use change shifts C-poor, productive ecosystems to C-rich, unproductive ecosystems and vice versa. J. Ecol., 91, 867-885.

FAO 2010. Greenhouse gas emissions from the dairy sector: A life cycle assessment. FAO, Rome, Italy, 97p.

FIL 2010. A common carbon footprint approach for dairy, Bulletin IDF, 46p.

Fontaine S., Bardoux G., Abbadie L. Mariotti A., 2004. Carbon input to soil may decrease soil carbon content. Ecol. Letters, 7, 314-320.

Gac A., Deltour L., Cariolle M., Dollé J.B., Espagnol S., Flénet F., Guingand N., Lagadec S., Le Gall A., Lellahi A., Malaval C., Ponchant P., Tailleur A., 2010a. GES’TIM, Guide méthodologique pour l’estimation des impacts des activités agricoles sur l’effet de serre. Version 1.2., Institut de l’Elevage, Paris, France, 156p.

Gac A., Dollé J.B., Le Gall A., Klumpp K., Tallec T., Mousset J., Eglin T., Bispo A., 2010 b. Le stockage de carbone par les prairies.

Institut de l’Elevage, Paris, France, 12p.

Gac A., Manneville V., Raison C., Charroin T., Ferrand M., 2010c. L’empreinte carbone des élevages d'herbivores : présentation de la méthodologie d’évaluation appliquée à des élevages spécialisés lait et viande. Renc. Rech.

Ruminants, 17, 335-342.

Gerber P., Vellinga T., Opio C., Steinfeld H., 2011. Productivity gains and greenhouse gas emissions intensity in dairy systems. Livest.

Sci., doi:10.1016.j.livsci.2011.03.012.

Haas G., Wetterich F., Köpke U., 2001.

Comparative intensive, extensified and organic grassland farming in southern Germany by process life cycle assessment. Agric. Ecosyst.

Environ., 83, 43-53.

Hegarty R., 2004. Genotype differences and their impact on digestive tract function of ruminants: a review. Austr. J. Expl. Agric., 44, 459467.

Henriksson M., Flysjö A., Cederberg C. and Swensson C., 2011. Variation in carbon footprint of milk due to management differences between Swedish dairy farms. Animal, 5, published online, doi:10.1017/S175173111 1000437.

Hindrichsen I.K., Wettstein H.R., Machmuller A., Kreuzer M., 2006. Methane emission, nutrient degradation and nitrogen turnover in dairy cows and their slurry at different milk production scenarios with and without concentrate supplementation. Agric.

Ecosyst. Environ., 113, 150-161.

Klumpp K., Soussana, J.F., Falcimagne R., 2007. Effects of past and current disturbance on carbon cycling in grassland mesocosms.

Agric. Ecosyst. Environ., 121, 59-73.

Kristensen T., Mogensen L., Knudsen M.T., Hermansen J.E., 2011. Effect of production system and farming strategy on greenhouse gas emissions from commercial dairy farms in a life cycle approach. Livest. Sci., 13p.

Ledgard S., Schils R., Eriksen J., Luo J., 2009. Environmental impacts of grazed clover/grass pastures. Ir. J. Agric. Food Res., 48, 209-226.

Le Gall A., Beguin E., Dollé J.B., Manneville V., Pflimlin A., 2009. Nouveaux compromis techniques pour concilier les impératifs d'efficacité économique et environnementale des systèmes d'élevage herbivore.

Proc. Journ. AFPF, 15p.

Leip A., Weiss F., Wassenaar T., Perez I., Fellmann T., Loudjani P., Tubiello F., Grandgirard D., Monni S., Biala K., 2010.

Evaluation of the livestock sector's contribution to the EU greenhouse gas emissions (GGELS), final report. Europ. Comm. Joint Res. Centre, 32p.

Loiseau P., Soussana J.F., Louault F., Delpy R., 2001. Soil N contributes to the oscillations of the white clover content in mixed swards under simulated grazing (Lolium perenne L, Trifolium repens L.). Grass Forage Sci., 56, 205-217.

Louault F., Pillar V.D., Aufrere J., Garnier E., Soussana J.F., 2005. Plant traits and function types in response to reduced disturbance in semi-natural grassland. J. Veget. Sci., 16, 151160.

Lovett D., Stack L., Lovell S., Callan J., Flynn B., Hawkins M., O’Mara F., 2005.

Manipulating enteric methane emissions and animal performance of late-lactation dairy cows through concentrate supplementation at pasture. J. Dairy Sci., 88, 2836-2842.

Lovett D.K., Shalloo L., Dillon P. and O’Mara F.P., 2006. A systems approach to quantify greenhouse gas fluxes from pastoral dairy production as affected by management regime. Agric. Syst., 88, 156-179.

Luo J., de Klein C.A.M., Ledgard S.F., Saggar S., 2010. Management options to reduce nitrous oxide emissions from intensively grazed pastures: A review. Agric. Ecosys.

Environ., 136, 282-291.

Macdonald K.A., Williams Y., Dobson-Hill B., 2010. Effectiveness of a nitrification inhibitor (DCn) on a Coastal Taranaki dairy

farm. Proc. N.Z. Grassland Assoc., 72, 147152.

Martin C., Morgavi D.P., Doreau M., 2009.

Methane mitigation in ruminants: from microbe to the farm scale. Animal, 4, 351-365.

Martin C., Pomiès D., Ferlay A., Eugène M., Martin B., Doreau M., Chilliard Y., 2010.

Methane output in dairy cows in response to long-term feeding of grass-based diets supplemented with linseed or rapeseed. McGeough E.J., McGinn S.M., (Eds). Banff, Canada, Proc.

Greenhouse Gases Anim. Agric., Conf., 134135.

Martin C., Pomiès D., Ferlay A., Rochette Y., Martin B., Chilliard Y., Morgavi D.P., Doreau M., 2011. Methane output and rumen microbiota in dairy cows in response to longterm supplementation with linseed or rapeseed of grass silage or pasture based diets. Proc.

N.Z. Soc. Anim. Prod., 71, 243-247.

Metz B., Davidson O.R., Bosch P.R., Dave R., Meyer L.A., 2007. Contribution of working group III to the fourth assessment report of the intergovernmental panel on climate change.

Cambr. Univ. Press, Cambridge, United Kingdom and New York, USA, 70p.

Monaghan R.M, Hedley M.J., Di H.J., McDowell R.W., Cameron K.C., Ledgard S., 2007. Nutrient management in New Zeland pastures. Recent developements and future issues. N.Z. J. Agric. Res., 50, 181-201.

Monaghan R.M, Smith L.C, Ledgard S.F., 2009. The effectiveness of a granular formulation of dicyandiamide (DCD) in limiting nitrate leaching from a grazed dairy pasture.

N.Z. J. Agric. Res., 2, 145-159.

Morgan J.A., Pataki D.E., Korner C., Clark H., Del Grosso S.J., Grunzweig J.M., Knapp A.K., Mosier A.R., Newton P.C.D., Niklaus P.A., Nippert J.B., Nowak R.S., Parton W.J., Polley H.W., Shaw M.R., 2004. Water relations in grassland and desert ecosystems exposed to elevated atmospheric CO2. Oecologia, 140, 11-25.

Nielsen P.H., Nielsen A.M., Weidema B.P., Dalgaard R., Halberg N., 2003. LCA food data base. http://www.lcafood.dk/ Pelletier N., Pirog R. and Rasmussen R., 2010. Comparative life cycle environmental impacts of three beef production strategies in the Upper Midwestern United States. Agric.

Syst., 103, 380-389.

Phetteplace H., Johnson D., Seidl A., 2001.

Greenhouse gas emissions from simulated beef and dairy livestock systems in the United States. Nut. Cycl. Agroecosys., 60, 99-102.

Reeder J.D., Schuman G.E., 2002. Influence of livestock grazing on C sequestration in semi-arid mixed-grass and short-grass rangelands. Environ Pollut., 116, 457-463.

Rossier D., Gaillard G., 2001. Bilan écologique de l’exploitation agricole, méthode et application à 50 entreprises. Office fédéral de l’agriculture, 163p.

Schils R.L.M., Verhagen A., Aarts H.F.M., Sebek L.B.J., 2005. A farm level approach to define successful mitigation strategies for GHG emissions from ruminant livestock systems. Nutr. Cycl. Agrosyst., 71, 163-175.

Schulze E.D., Luyssaert S., Ciais P., Freibauer A., Janssens I.A., Soussana J.F., Smith P., Grace J., Levin I., Thiruchittampalam B., Heimann M., Dolman A.J., Valentini R., Bousquet P., Peylin P., Peters W., Rodenbeck

INRA Productions Animales, 2011, numéro 5

430 / J.-B. DOLLE et al

C., Etiope G., Vuichard N., Wattenbach M., Nabuurs G.J., Poussi Z., Nieschulze J., Gash J.H., 2009. Importance of methane and nitrous oxide for Europe’s terrestrial greenhouse-gas balance. Nature Geosci., 2, 842-850.

Smith P., Chapman S.J., Scott W.A., Black H.I.J., Wattenbach M., Milne R., Campbell C.D., Lilly A., Ostle N., Levy P.E., Lumsdon D.G., Millard P., Towers W., Zaehle Z., Smith J.U., 2007. Climate change cannot be entirely responsible for soil carbon loss observed in England and Wales, 1978-2003. Global Change Biol., 13, 2605-2609.

Soussana J.F and Hartwig U.A., 1996. The effects of elevated CO2 on symbiotic N2 fixation : a link between the carbon and nitrogen cycles in grassland ecosystems. Plant and Soil, 187, 321-332.

Soussana J.F., Allard V., Pilegaard K., Ambus P., Amman C., Campbell C., Ceschia E., Clifton-Brown J., Czobel S., Domingues R., Flechard C., Fuhrer J., Hensen A., Horvath L., Jones M., Kaspe G., Martin C., Nagy Z., Neftel A., Raschi A., Baronti S., Rees R.M., Skiba U., Stefani P., Manca G., Sutton M., Tuba Z., Valentini R., 2007. Full accounting of the greenhouse gas (CO2, N2O, CH4) budget of nine European grassland sites. Agric. Ecosyst.

Environ., 121, 121-134.

Soussana J.F., Lüscher A., 2007. Temperate grasslands and global atmospheric change: a review. Grass Forage Sci., 62,127-134.

Soussana J.F., Tallec T., Blanfort V., 2010.

Mitigating the greenhouse gas balance of ruminant production systems through carbon sequestration in grasslands. Animal, 4, 334350.

Tomasula P.M., Nutter D.W., 2011.

Mitigation of greenhouse gas emissions in the production of fluid milk. Food Nutr. Res., 62, 42-88.

Tubiello F., Soussana J.F., Howden S.M., Easterling W., 2007. Crop and pasture response to climate change. Proc. Nat. Acad. Sci. U.S.A, 104, 19686-19690.

Van der Werf H.M.G., Kanyarushoki C., Corson M.S., 2009. An operational method for the evaluation of resource use and environmental impacts of dairy farms by life cycle assessment. J. Environ. Manage., 90, 3643-3652.

Vellinga T, de Haan M.H.A., Schils R.L.M., Evers A., van den Pol-van Dasselaar A., 2011.

Implementation of GHG mitigation on intensive dairy farms: farmers preferences and variation in cost effectiveness. Livest. Sci., 137, 185-195.

Vérité, R., Faverdin, P., Agabriel, J., 1997. In: Recent advances in animal nutrition. Wiseman J., Garnsworthy P.C. (Eds). Nottingham Univ.

Press, Nottingham, UK, 153-166.

Veysset P., Lherm M., Bébin D., 2010.

Energy consumption, greenhouse gas emissions and economic performance assessments in French Charolais suckler cattle farms : Model-based analysis and forecasts. Elsevier Agric. Syst., 103, 41-50.

Veysset P., Lherm M., Bébin D., 2011.

Productive, environmental and economic performances assessments of organic and conventional suckler cattle farming systems. Org. Agric., 1-16.

Watson R., Noble I., Bolin B., Ravindranath N., Verardo D., Dokken D., 2002. Land Use, Land Use Change, and Forestry. Camb. Univ.

Press, Cambridge, UK. Woodward S.L., Waghorn G.C. and Thomson N.A., 2006. Supplementing dairy cows with oils to improve performance and reduce methane – does it work ? Proc. N.Z. Soc. Anim. Prod., 66, 176-181.

Zehetmeier M., Baudracco J., Hoffmann H., HeiBenhuber A., 2011. Does increasing milk yield per cow reduce greenhouse gas emissions ? A system approach. Animal, 13p.

Résumé Le contexte environnemental actuel, tant politique (objectifs de réduction des émissions de gaz à effet de serre) que sociétal (information du consommateur), nécessite de préciser les impacts de l'activité d'élevage bovin en matière de changement climatique. L'enjeu est de connaître précisément les niveaux d'émissions de gaz à effet de serre (GES) et de stockage de carbone, des différents modes de production. Pour cela, une évaluation basée sur la méthodologie de l'Analyse du Cycle de Vie (ACV) est mise au point à l'échelle du système d'élevage. Cette approche permet d'avoir une vision globale de l'activité d'élevage intégrant l'ensemble des processus internes et externes au fonctionnement de l'exploitation. Ainsi pour les systèmes laitiers français, l'empreinte carbone brute du lait est en moyenne de 1,26 kg CO2/kg de lait. La prise en compte du stockage de carbone sous les prairies et les haies se traduit par une compensation comprise entre 6 et 43% selon les systèmes, en fonction de la part de prairies. L'empreinte carbone nette du lait français est alors en moyenne de 1,0 kg CO2/kg de lait. Dans les systèmes bovins viande français, l'empreinte carbone brute est comprise entre 14,8 et 16,5 kg CO2/kg viande vive en fonction du système de production (naisseur vs naisseur/engraisseur). Après prise en compte du stockage de carbone qui permet une compensation comprise entre 24 et 53%, l'empreinte carbone nette est comprise entre 7,9 et 11,3 kg CO2/kg viande vive. De nombreux leviers d'action sont identifiés dans les systèmes d'élevage de ruminants pour réduire l'empreinte carbone des produits au portail de la ferme. Certains concernent une optimisation des systèmes de production (ajustement des apports alimentaires, gestion de la fertilisation…) et se traduisent par des économies en matière d'intrants. D'autres nécessitent la mise en place de nouvelles technologies et se traduiront donc par un investissement ou un coût de fonctionnement supérieur aux schémas actuels de production.

Abstract Greenhouse gases in cattle breeding : evaluation and mitigation strategies In today's environmental context, as much political (the reduction of greenhouse gas emissions) as social (consumer demands for information concerning food products), there is a need to determine the influence of ruminant livestock on climate change. It has become crucial to quantify precisely the levels of greenhouse gas emissions (GHG) and carbon sequestration for different ruminant livestock systems by using Life Cycle Assessment (LCA) to calculate the carbon footprint of dairy and beef farms. This approach allows us to account for GHG in relation with direct and indirect emissions. Thus in the French dairy system, the average gross carbon footprint is 1.26 kg CO2 /kg milk. Carbon sequestration under grassland and hedges compensates for GHG emissions ranging from 6 to 43% according to the system utilized. Consequently the net carbon footprint is 1.0 kg CO2 / kg milk. In the French beef system, the gross carbon footprint for raw materials is comprised between 14.8 and 16.5 kg CO2 / kg of live meat depending on the production system. It must be mentioned that carbon sequestration represents compensation in a range of 24 to 53% and the net carbon footprint is between 7.9 and 11.3 kg CO2 / kg of live meat. These investigations demonstrate that numerous mitigation actions

INRA Productions Animales, 2011, numéro 5

Greenhouse gases in cattle breeding: evaluation and mitigation strategies / 431

have been identified in the livestock systems to reduce the carbon footprint of milk and meat at the farm gate. Some of them concern management practices (adjustment of dietary intake, fertilization management‌) which result in substantial savings in agricultural expenses. Others require the installation of new technologies which would require additional funds to improve the production processes.

DOLLE J.-B., AGABRIEL J., PEYRAUD J.-L., FAVERDIN P., MANNEVILLE V., RAISON C., GAC A., LE GALL A., 2011. Greenhouse gases in cattle breeding: evaluation and mitigation strategies. Doreau M., Baumont R., Perez J.M. (Eds).

Dossier, INRA Prod. Anim. 24, 415-432.

INRA Productions Animales, 2011, numĂŠro 5