If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Enteric fermentation from dairy cows is a major source of methane. Significantly and rapidly reducing those emissions would be a powerful lever to mitigate climate change. For a given productivity level, introducing fodder with high sources of omega 3 content such as grass or linseed in the feed ration of dairy cows both improves the milk nutritional profile and reduces enteric methane emissions per liter. Changing cows' diet may represent additional costs for dairy farmers and calls for the implementation of payments for environmental services to support the transition. This paper analyzes 2 design elements affecting the effectiveness of a payment conditioned to the reduction of enteric methane emissions: (i) the choice of emission indicator capturing the effect of farmers' practices (ii) the payment amount relative to the extra milk production costs incurred. Using representative farm-level economic data from the French farm accountancy data network, we compare enteric methane emissions per liter of milk calculated with an Intergovernmental Panel on Climate Change Tier 2 method, to baseline emissions from a Tier 3 method accounting for diet effects. We also quantify the extra milk production costs of integrating more grass in the fodder systems by estimating variable cost functions for different dairy systems in France. Our results show the relevance of using an emission indicator sensitive to diet effects, and that the significance and direction of the extra-costs for producing milk with more grass differ according to the production basin and the current share of grasslands in the fodder crop rotation. We stress the importance of developing payments for environmental services with well-defined environmental indicators accounting for the technical problem addressed, and the need to better characterize heterogeneous funding requirements for supporting a large-scale adoption of more environment-friendly practices by farmers.
INTERPRETIVE SUMMARY: This study provides insights on the design of effective payments for environmental services conditioned to enteric methane emissions reduction. Two indicators of enteric emissions were compared, and the extra production costs for adding more grass in fodder systems as an emission reduction lever were computed for representative French medium and large dairy farms. Results show that an emission indicator accounting for diet effects is crucial to accurately monitor farmers' effort. Moreover, farms face heterogeneous extra-costs for changing cows' feed ration, and it should be taken into account in the definition of the payment amount to ensure a large-scale adoption.
INTRODUCTION
The agricultural sector is a major source of greenhouse gases (GHG), accounting for 10% of EU-KPs (European Union, United Kingdom and Iceland) in 2018 (
Inventaire des émissions de polluants atmosphériques et de gaz à effet de serre en France - Format Secten. National inventory of GHG and air pollutant emissions in France (in French).
). Methane is the second contributor to radiative forcing. Currently, the global warming potential of methane is set at 28 times higher than the global warming potential of carbon dioxide over 100 years and 84 over 20 years (
in: Stocker T.F. Qin D. Plattner G.-K. Tignor M. Allen S.K. Boschung J. Nauels A. Xia Y. Bex V. Midgley P.M. Climate Change 2013 the Physical Science Basis: Working Group I Contribution to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press,
Cambridge, New York2013: 659-740
). As methane is a short-lived climate pollutant continuously destroyed in the atmosphere, its effect on climate change depends mostly on short-term emission rates. In theory, decreasing the methane emissions rate below its natural destruction rate would have a cooling effect (
). Therefore, a significant reduction in methane emissions, in particular from agricultural activities, would rapidly mitigate climate change and is a powerful lever to meet the European Union's 2050 climate targets (
). The quantity of methane produced during the digestive process of ruminants depends highly on the characteristics of the animal itself, such as the breed, body weight and age (
). However, enteric methane emissions are also directly related to farming practices, in particular the amount of feed intake and composition, and the proportion of carbohydrates that feed ration contains (
). In particular, for a given productivity level, enteric methane emissions decline as dairy cow feed is enriched with α-linolenic acid ALA (polyunsaturated fatty acid of the omega-3 family), for which the main natural sources are linseed and grass fodders (
Methane output and rumen microbiota in dairy cows in response to long-term supplementation with linseed or rapeseed of grass silage or pasture based diets.
in: Proceedings of the New Zealand Society of Animal Production. volume 71. New Zealand Society of Animal Production,
Invercargill2011: 243-247
). To accurately monitor the evolution of enteric methane emissions on dairy farms, one must consider both dimensions (productivity and feeding). Authors show enteric methane emissions can differ significantly from one indicator to another and recommend considering both production intensity and feed usage for more accurate estimates (
Changing management practices, and in particular cows diet, to decrease enteric methane emissions may be costly for many farmers. Economic incentives are currently being developed in the agricultural sector to support the reduction of GHG emissions. While studies show taxes and emission permits are the most efficient mitigation policy options, they are unpopular tools among producers and consumers (
). Alternatively, there is growing interest in supporting financially changes of management practices through payments for environmental services (PES), as more acceptable tools on the short-term. Targeting enteric methane emissions is of main interest for potential contributors, due to the high and rapid mitigation potential. In particular, supporting diet change levers such as increasing grass use, is also highlighted by life cycle assessment approaches has having the potential to reduce the total global warming potential and biodiversity damage of milk production (
). Developing PES schemes for the reduction of enteric methane emissions raises the question of the choice of a practical emission indicator sensitive to both diet and productivity effects, easily applicable and measurable on farm. While numerous other motivations may encourage farmers to join a PES program, such as improving milk quality, environmental quality, zootechnical performance and the image of agriculture, economic interests are likely to be crucial factors. An efficient payment level of a PES scheme targeting GHG emissions is equal to the socially optimal carbon price. In the EU, carbon tax levels differ largely among countries, illustrating the difficulty in pricing GHG emissions and therefore the social value of mitigation efforts. In 2021, it was set at 120€/teqCO2 in Sweden and 45€/teqCO2 in France, while the EU ETS (European Trading System) market price was 44€/teqCO2 (
In this paper, we aim to provide insights into the design of PES schemes targeting the reduction of enteric methane emissions in dairy farms by examining 2 aspects of a payment mechanism for which failing to consider the feeding dimension could undermine its effectiveness in cutting enteric methane emissions: the choice of emissions indicator defining the environmental service, and the level of payment. By comparing an indicator constructed using a methodology applied in a real-life PES case study from France to the Tier 2 indicator currently used in the French annual GHG emissions inventory which considers productivity only, we examine how the diet of dairy cows influences the enteric methane emissions attributed to farms. Changing the diet of cows to improve the milk fatty acid profile can generate additional production costs that are not yet evaluated. The second contribution is to quantify the additional cost of a change in the diet of cows at the farm level to evaluate the economic incentives needed for improving dairy systems toward more environment-friendly practices. We estimate the variable cost function of dairy farms at the scale of France and for different fodder systems.
BACKGROUND ON ENTERIC METHANE EMISSION INDICATORS
Many indicators of enteric methane emissions have been developed and adapted to specific constraints, often related to the scale of their application and the data available for estimating emissions (
Invited review: Large-scale indirect measurements for enteric methane emissions in dairy cattle: A review of proxies and their potential for use in management and breeding decisions.
). For result-based PES, emissions should regularly and easily be monitored on farm. The emission indicator chosen must therefore be easily measurable by the farmers and/or the paying agent using a simple methodology, representative of the environmental target and reliable (based on strong and reliable scientific evidence) (
Biodiversity protection through results based remuneration of ecological achievement. Report Prepared for the European Commission, DG Environment, Contract No ENV.B.2/ETU/2013/0046.
Institute for European Environmental Policy,
London2014
). In addition, it must be sensitive to the different dimensions on which farmers can act to decrease emissions, in particular cows' diet and productivity.
The Intergovernmental Panel on Climate Change (IPCC) defines 3 families of methods, to be applied for national inventories according to data availability. Tier 1 methods attribute default yearly enteric methane emissions factor per dairy cow. Tier 1 methods provide aggregate estimates, and are not adequate for monitoring changes over time and taking into account for the variability of dairy farming practices. Tier 2 methods improve the accuracy of emission factors by including feed intake estimates of a representative diet and dairy cow (
). Finally, Tier 3 methods require a precise characterization of cows' diet to account for digestibility. Both Tier 2 and Tier 3 approaches for France are presented in
applied a Tier 1 enteric methane emission indicator at the micro scale, but as part of an overall assessment of the relative GHG emissions and economic performance of farms. Life cycle assessment analyzes calculating the carbon footprint of dairy products tend to apply Tier 2 methods (
). Bioeconomic models estimating emissions abatement costs in the agricultural sector use Tier 3 indicators of enteric methane emissions precise enough to capture both productivity and diet effects (
). Most Tier 3 and “individualized” Tier 2 approaches mentioned above require a large amount of detailed individual farm-level data on cows' feed composition, making it too costly to be applied as monitoring indicators on a large amount of farms. However, some of the Tier 3 methods can use information from milk analyzes and thus, be more easily integrated in farm routines.
A TIER 3 INDICATOR APPLIED IN THE FRENCH ECO-METHANE RESULT-BASED SCHEME
Numerous studies have been carried out to understand the connection between dairy cows' enteric methane emissions, fat intake and milk composition (
Methane output and rumen microbiota in dairy cows in response to long-term supplementation with linseed or rapeseed of grass silage or pasture based diets.
in: Proceedings of the New Zealand Society of Animal Production. volume 71. New Zealand Society of Animal Production,
Invercargill2011: 243-247
). In particular, experimental research shows that enteric methane emissions in gCH4/L (Methane) can be calculated from milk productivity in kg/cow/year (Productivity) and the ratio of the sum of fatty acids with 16 carbon atoms or less (FA ≤ C16) to the total amount of fatty acids (TotalFA) in the milk they produce (1). This ratio has a strong biological causal relationship with methanogenesis in the rumen and is significantly reduced by more sources of omega-3 in cows' diet.
[1]
This formula was coinvented by teams from the animal feed manufacturing company Valorex (P. Weill and G. Chesneau) and the French National Institute for Agricultural Research (INRA) (Y. Chilliard, M. Doreau and C. Martin), and received a patent under the title “Method for evaluating the quantity of methane produced by a dairy ruminant and method for decreasing and controlling such quantity” (WO2009156453A1) (
Weill, P., G. Chesneau, Y. Chilliard, M. Doreau, and C. Martin. 2009. Method to evaluate the quantity of methane produced by a dairy ruminant and method for decreasing and controlling this quantity. Valorex, assignee. Pat. No. WO 2009/156453 A1.
). The necessary data to implement it are relatively easy to collect on farm. One needs milk productivity and an analysis of the fatty acid composition of milk, obtained with infrared spectroscopy. Milk infrared spectroscopy is relatively simple to integrate into the milk analyzes routines of dairy farms and involves low costs. Since dairy farms already undertake milk analyzes on a regular basis, it makes it a relatively cheap indicator to monitor emissions. In addition, the equation calculates enteric methane emissions per unit of product by taking into account both milk productivity and feed quality, hence captures both dimensions on which farmers can dedicate effort to reduce emissions.
The indicator described above is used since 2011 to evaluate the reduction of enteric methane emissions of dairy farms participating in the PES scheme Eco-Methane in France, implemented and coordinated by the Bleu-Blanc-Coeur association (BBC). Eco-Methane meets the reference PES definition by
. In the scheme, private stakeholders (service users) give financial support to volunteer dairy farmers (service providers) for actions that contribute to climate change mitigation (environmental services). The payment is conditional and proportional to the reduction of CO2eq, making the scheme a result-based PES. The reduction is calculated relative to baseline monthly emission levels attributed to the farm that are representative of its type of fodder system. Through the definition of different baseline scenarios, the PES design partially considers the variability in the potentiality of environmental services provision according to the production basin and the fodder system. Hence, rather than rewarding farms that produce the least emissions per unit of product (which would tend to favor the most productive farms), Eco-Methane supports all emission reduction efforts.
The Eco-Methane scheme uses 11 scenarios of baseline emission levels representative of 11 fodder systems of French specialized dairy farms. These fodder systems were characterized by the French Dairy Interbranch Organization (CNIEL) in collaboration with the French Livestock Institute (IDELE) in 2009, based on large production basins (plain and mountainous areas), and the proportion of corn in fodder crop rotation systems (
Observatoire de l'alimentation des vaches laitières: description des 8 principaux systèmes d'élevage. Edition 2015–2018. Dairy cows' feed observatory: description of the 8 main farming systems. 2015–2018 edition (in French).
Démarche Environnementale : La Démarche Éco-Méthane de Bleu-Blanc-Cœur. Environmental Engagement: The Eco-Methane Programme of Bleu-Blanc-Coeur (in French).
). BBC pays farmers according to their reduction of methane emissions in CO2eq with a financial envelope made of donations from private actors (15€/tCO2eq on average in 2017). The main strengths of the scheme lie in the strong scientific foundations of the method for quantifying emissions and the easy participation procedure for dairy farmers. Each contract signatory commits to provide a monthly milk analysis to the association and to include feed with a high content of sources of omega-3 in dairy cow rations (alfalfa, extruded linseed, grass). This data is used to estimate enteric methane emissions using (1). Eco-Methane is recognized by the United Nations as a GHG emission reduction project eligible for issuing carbon credits (
Joint Implementation Project FR1000365: “Réduction des émissions de méthane d'origine digestive par l'apport dans l'alimentation des vaches laitières de sources naturelles en Acide Alpha Linolénique (ALA)” (in French).
Observations of a balanced panel of 735 French farms specialized in dairy milk production (OTEXE 45) from France's farm accountancy data network (FADN) for the years 2016 to 2018 was selected for the analysis. This European database is freely accessible online and representative of socioeconomic and accountancy information of medium and large dairy farms contributing to more than to more than 90% of the gross production and utilized agricultural area. Due to this characteristic, it is a particularly relevant data set to investigate emission abatement costs at the national level. As the compositions of the feed ration and milk are not surveyed, information on the diet of dairy cows is limited and in particular the fatty acid profile is unknown. Instead, data on the fodder crop rotation systems are used to assess a change in crop rotation and approximate a change in feed composition. Descriptive statistics of the sample are presented (Table 1).
Table 1Description of the sample (2205 observations)
Information on surfaces and the number of dairy cows available in the database are ranges of values. We constructed the variables used in the analysis by taking the lower value of the range for each observation. Source: The authors, based on French FADN data.
Information on surfaces and the number of dairy cows available in the database are ranges of values. We constructed the variables used in the analysis by taking the lower value of the range for each observation. Source: The authors, based on French FADN data.
Information on surfaces and the number of dairy cows available in the database are ranges of values. We constructed the variables used in the analysis by taking the lower value of the range for each observation. Source: The authors, based on French FADN data.
Information on surfaces and the number of dairy cows available in the database are ranges of values. We constructed the variables used in the analysis by taking the lower value of the range for each observation. Source: The authors, based on French FADN data.
Information on surfaces and the number of dairy cows available in the database are ranges of values. We constructed the variables used in the analysis by taking the lower value of the range for each observation. Source: The authors, based on French FADN data.
35
55
58
70
Agricultural Work Unit
1.0
2.0
1.8
2.1
Purchase of cattle feed concentrates (€)
14,326.0
24,996.5
32,853.2
43,645.0
1 Information on surfaces and the number of dairy cows available in the database are ranges of values. We constructed the variables used in the analysis by taking the lower value of the range for each observation.Source: The authors, based on French FADN data.
We use the Tier 2 method used in the French annual inventory of GHG emissions to define an indicator of enteric methane emissions sensitive to productivity effects. Dairy cows emission factors are calculated from equation (2) (
Organisation et méthodes des inventaires nationaux des émissions atmosphériques en France OMINEA - 17 ème édition. Inventory methodology report of GHG and air pollutant emissions in France - 17th edition (in French).
). The parameters of the formula are such that they are representative of breeding, feeding and genetic conditions in France. Based on this calibration, the emission factor per cow varies according to milk productivity.
(2)
The emission factor EF (kgCH4/cow/year) can be easily calculated for each farm of the FADN from the milk production (kg/year) of the herd (Milk production) and the number of dairy cows (Ncows). We then derive an emission indicator (TIER2) per liter of milk (gCH4/L), capturing variability according to milk productivity (L/cow/year) (3).
(3)
Due to the absence of data on dairy cows' diet, a Tier 3 method cannot be applied to evaluate individual emissions of FADN farms and capture both productivity and diet effects. In particular, data are too limited to estimate individual enteric methane emissions of French farms using (1). They are, however, sufficient to identify their Eco-Methane scenario and therefore the baseline emissions corresponding to their fodder system. Baseline emissions from the 11 scenarios are available per month and were obtained from BBC. We calculate the annual average to define the Eco-Methane baseline emissions indicator (Eco-Methane baseline) (Table 2).
Table 2Characteristics of the eleven baseline scenarios used in the Eco-Methane scheme
Scenario
Corn in the fodder area
Production basin
Eco-Methane baseline (gCH4/L)
1
More than 30%
Plains outside the western region
15.75
2
Plains of the western region
15.92
3
Between 10 and 30%
Plains outside the western region
15.83
4
Plains of the western region
16.43
5
Less than 10%
Plains outside the western region
16.56
6
Plains of the western region
17.38
7
More than 10%
Mountains
15.96
8
Less than 10%
Mountains of the Massif Central
17.13
9
Mountains of the Northern Alps
17.83
10
Mountains of Franche-Comté
16.22
11
Other mountains
17.20
Source: The authors, based on BBC data.
Source: The authors, based on French FADN and BBC data.
An individual baseline scenario was assigned to each farm of the sample based on 2 criteria: the location and the share of corn silage in the fodder area of the farm. In the FADN database, the farm location variable corresponds to the 21 old French administrative regions (the administrative divisions were changed in 2015), while the Eco-Methane scenarios are defined according to large production basins built from a lower administrative level (departments). It was therefore necessary to allocate a production basin to each administrative region. For the regions with departments belonging to different production basins, we allocated the basin of the departments producing the highest volumes of milk to the entire region. This attribution was made using the 2018 annual dairy survey (
Estimation of the extra-cost of milk production with more grass in the fodder system
In France, the closest financial tool to a carbon tax is the Climate and Energy Contribution proportional to the carbon dioxide content of energy products (fossil fuels) (
La Contribution Climat Energie en France : fonctionnement, revenus et exonérations. The Climate-Energy Contribution in France: operation, revenues and exemptions.
). The contribution level was 30€/tCO2eq in 2017 and increased to 40€/tCO2eq in 2018 and 2019. Farmers participating in Eco-Methane received an average of 15€/tCO2eq in 2017, suggesting that the payment of the scheme is suboptimal and provides little incentive to participate (
Démarche Environnementale : La Démarche Éco-Méthane de Bleu-Blanc-Cœur. Environmental Engagement: The Eco-Methane Programme of Bleu-Blanc-Coeur (in French).
). Evaluating farmers' willingness to accept is necessary to define a more efficient price and trigger a large scale adoption of practices decreasing enteric emissions.
Since the composition of the feed ration and milk of cows are not available in the FADN, the effect of an improvement of the fatty acid profile on milk production costs cannot be analyzed directly. Instead, an evolution of the fodder crop rotation is assumed. As grass is a high source of omega-3 fatty acids strongly encouraged in Eco-Methane, we assume that a commitment to the program would lead to an increase in grassland surfaces in farms. This hypothesis is quite strong and implies that the estimation of extra costs does not consider either the strategy of supplementing the ration with other feeds with high omega-3 content such as extruded linseed or the optimization of grazing increasing grass yield and quality without necessarily increasing grassland surfaces.
in: Fuss M. McFadden D. Production Economics: A Dual Approach to Theory and Applications, Volume I: The Theory of Production. North Holland,
Amsterdam1978: 2-109
), we estimate a variable cost function describing expenditures in variable production factors x with exogenous input prices w that minimize variable costs given the production level y targeted by the farmer and available quasifixed inputs z such as land, labor and equipment that are assumed to be predetermined in the short term. The cost minimization approach is motivated by the fact that French dairy farmers are constrained in the quantity of milk they produce in the terms of their contract with dairy plants (
). It is confirmed in our data (see Table A2 in the Appendix), as we observe a low variability of milk volumes across years for a same farm. The variable cost function (4) meets theoretical properties that we empirically check.
(4)
It must be concave, nondecreasing and homogeneous of degree 1 in variable input prices, decreasing with (binding) fixed factors of production, and monotonic according to output levels.
We estimate a system of equations comprising a homogeneous translog cost function (5) in which variable costs VC correspond to intermediate consumption and the variable input cost share functions (6) and (7), derived from Shephard's lemma. The translog functional form is commonly used in the literature on the cost structure and efficiency of dairy farms because of its flexibility and the possibility of imposing homogeneity of degree 1 (
). i and t are indices for individuals and years, respectively. We assume that dairy farms produce one output, the quantity Y1 of the milk of cows produced per year (milk production). Considering the joint production of milk and a bad output (enteric methane emissions) would be of main interest. This is the approach used in (
) for instance. Implementing it in our study would require having enough information to compute an individualised enteric methane emission indicator capturing both cows' productivity and diet. The FADN data set does not provide enough information on cows' diet to compute the Eco-Methane indicator for each individual farm and we cannot precisely compute the “bad” output.
We consider 2 variable inputs, fuel X1 and cattle feeding stuffs X2, for which the expenses represent a high share of intermediate consumption. The choice of including fuel is motivated by the possibility of calculating farm-level fuel prices and therefore capturing more heterogeneity. The price of fuel W1 (fuel price) is calculated from the nonroad gas oil expenses and volumes. As individual cattle feeding stuff prices are not available in the data, W2 (feed price) is measured by the index of purchase prices of the means of agricultural production (IPAMPA) for adult cattle feeding stuffs of year t-1, available for each French current administrative region. While in practice, a bargaining effect may create endogeneity between milk production and feed prices, we assume it is sufficiently low not to overestimate too much milk marginal production costs (
). Marginal costs tend to be overestimated when the adjustment of the feed unit prices to a change of farm size is not controlled for. In our case, capturing this endogeneity issue would require individual farm feed price data, information we do not have.
found evidence of a price bargaining effect for hog production, but lower than the technology effect. For hog production, feed represents a larger share of the costs (about 60% of total cost) than for dairy production (25% of variable cost, see Table A2 of the appendix). Therefore, the price bargaining effect is expected to be lower for dairy farms.
We also include 3 quasifixed inputs. Grassland surface Z1 (grassland) includes permanent and temporary pastures, alfalfa for dehydration and other artificial fodders. We add 2 other quasifixed factors of production: machinery and constructions fixed assets Z2 (capital) and annual work units Z3 (labor). Since sample farms are all specialized in dairy milk production, we do not consider them as multi-output farms. We verify the production of crops and other livestock products represents a much lower share of total gross products (see Table A2 in the Appendix). Yet, we add the aggregated volume Y2 of the other products of the farm (other productions) as a control variable to capture heterogeneity linked to diversification. Y2 is calculated as the total gross product of the year (crop products, livestock products and other products) net of animal purchases and milk production of the cows, deflated by the French agricultural producer price index (API) of year t. We further control for the farm total utilized agricultural area Q1 (utilized agricultural area), 20 regional dummies Dri approximating the pedoclimatic conditions, and time fixed effects with 2 yearly dummies Tt. The coefficients α, β, δ, ν, ρ, ζ, θ, γ, and µ are the parameters to be estimated and describe the effects of the covariates on the independent variables. eit, u1it and u2it are the error terms assumed to follow a normal distribution.
(5)
(6)
(7)
To account for the correlations between the error terms of the different equations, we use a 3-stage least squares regression analysis. The system of equations (5) + (6) + (7) is estimated for all the farms of the sample and then for the 3 major production basins and groups of Eco-Methane scenarios defined in Table 2 to identify potential differences in extra costs according to the type of dairy system. The different steps to derive the specification of the system are described in the Appendix. Descriptive statistics of the model variables of the sample and subsamples are presented in Table 3.
Table 3Mean values of the model variables (standard deviation in parentheses)
First, we present the allocation of the FADN dairy farms in the Eco methane scenarios, then we compare the 2 indicators Tier 2 and the Eco-Methane one. Finally, we discussed the extra-costs associated with an increase in farm grassland area based on estimated marginal costs.
Allocation of Eco-Methane scenarios and distinction of 3 milk production basins
Following the allocation of the Eco-Methane baseline scenarios to the farms of the sample, their proportion within each former administrative region can be observed (Figure 1). Farms in the regions of the western plains basin (Brittany, Pays de la Loire and Basse-Normandie) represent 44% of the sample and are characterized by a strong dominance of corn silage and few grasslands in forage crop rotation systems. For example, 68% of farms in Brittany would be assigned scenario 2 with more than 30% of corn in the forage crop rotation and 28% scenario 4 with 10 to 30%. The administrative regions of the plain production basin outside the western region (31% of the sample) and of the mountainous areas (24% of the sample) are quite different. Some plain regions, such as Rhône-Alpes, contain a high proportion of grazing systems dominated by grasslands (less than 10% of corn silage in the fodder crop rotation), while others, such as the Centre, have more intensive systems dominated by corn silage. All the observations from Languedoc-Roussillon correspond to grazing systems, while the observations from Midi-Pyrénées have a relatively small proportion (32%). Due to the missing information on departments in the FADN data set and our scenario allocation procedure, some farms have been allocated to a plain system scenario, while, in reality, they are located in a mountainous department and vice versa. This allocation might partly explain the large share of farms with grazing systems in Rhône-Alpes. Nevertheless, those farms produce relatively low volumes of milk in comparison with plain farms from the same region.
Figure 1Distribution of Eco-Methane baseline scenarios among French old administrative regions.
Enteric emissions: relation to productivity and fodder system
The mean of the Tier 2 indicator for the sample is 18.5 gCH4/L, while it is 16.3 gCH4/L for the Eco-Methane baseline indicator, suggesting that considering the fodder cropping system in the calculation revises enteric emissions downwards. Both indicators show a decrease in emissions per liter of milk as productivity increases (Table 4). Dairy systems with lower enteric emissions per liter of milk are productive systems with more than 30% of corn silage in the fodder crop rotation, while grass-based systems have the highest emissions. The comparison of the 2 indicators allows to nuance the performance of the different systems in terms of diet and productivity effects.
Table 4Average enteric emissions of the sample according to the two indicators
Scenarios
Plains outside the western region
Western plains
Mountains
1, 3, 5
1
3
5
2, 4, 6
2
4
6
7–11
7
8 to 11
% corn in fodder area
>30
30–10
<10
>30
30–10
<10
≥10
<10
Sample share (%)
31.4
10.0
9.9
11.5
44.3
23.7
16.8
3.8
24.3
6.5
17.8
Productivity (1,000L/cow)
6.72
7.65
6.94
5.72
6.98
7.33
6.79
5.59
6.20
6.91
5.94
TIER2 (gCH4/L)
18.55d
17.46l
18.22fhj
19.78a
18.23ehi
17.80k
18.38defg
20.22a
19.07c
18.19gij
19.40b
Eco-Methane baseline (gCH4/L)
16.07q
15.75u
15.83t
16.56n
16.24p
15.92s
16.43°
17.38l
16.55n
15.96r
16.76m
Mean values with different superscripts differ (p-value <0.05).
Source: The authors, based on French FADN and BBC data.
According to indicator Tier 2, farms in mountains emit significantly more methane per liter of milk on average than farms in the plains which can be explained by their lower productivity. The same observation is made with the Eco-Methane indicator, but the difference between the groups is less (Table 4 and Figure 2). This finding shows that a diet effect partly compensates high emissions from low productivity in mountainous areas. The average share of grass in mountain fodder crop rotation systems is 90%, providing a diet rich in sources of omega-3 to dairy cows and contributing to the mitigation of enteric emissions. In fact, we observe that the difference between the 2 indicators is large for low productivity and grass-based dairy systems, and low for productive systems with a large share of corn silage in the fodder area (see Figure A1 in the Appendix).
Figure 2Average enteric emissions in French old administrative regions according to the 2 indicators Source: The authors, based on French FADN and BBC data.
In plain areas, the average methane emissions per liter of milk of dairy farms with a high share of corn silage in the fodder area are higher in western regions because of lower productivity (Table 4). However, the difference between the 2 plain production basins is not significantly different for a fodder system with an intermediate share (scenario 3 vs 4), or a low share (scenario 5 vs 6) of corn silage. Therefore, computing enteric emissions with the Tier 2 indicator results in no significant difference between the 2 plain production basins for a same fodder system. However, the Eco-Methane baseline attributed to plain farms outside the western region is lower. In western regions, farms have a significantly lower share of grass in the fodder rotation system (P-value < 0.01). It suggests that cows receive less omega-3 rich fodder in their ration.
Our observations are in line with the literature (
). A recent meta-analysis of life cycle assessments also highlighted the negative relationships between milk yield and enteric methane emissions on the one hand and pasture intake and enteric methane emissions on the other hand (
). Choosing an adequate indicator of enteric methane emissions is the topic of ongoing debate. This analysis illustrates the importance of the choice of environmental indicator when designing a PES scheme targeting the reduction of enteric methane emissions in dairy farms, as it is likely to affect its environmental performance. We show the relevance of using an emission indicator sensitive to diet when defining economic incentives for the reduction of GHG emissions.
Detailed indicators are the most cost-effective, but due to heavy data collection needs (precise feed digestibility and composition), this advantage decreases when applied at a large scale (
). An indicator such as the one used in Eco-Methane presents several advantages to be implemented on a large scale, and be a better proxy compared with the most widely used Tier 2 and data demanding Tier 3 indicators. It is precise enough to capture the efforts of farmers on both cow productivity and feed ration composition, and it considers the potential of an omega-3-rich diet as a climate change mitigation practice. This feeding strategy is already implemented in dairy systems integrating a large share of grasslands in their fodder crop rotation systems, with the side provision of other environmental benefits (biodiversity maintenance). Additionally, the fatty acid composition of milk provides information on the complementary health benefits for consumers of an increase in sources of omega-3 fatty acids in the diet of dairy cows (
). The accuracy of indicators based on milk analyzes could be further improved by controlling for factors likely to affect the correlation between milk fatty acid composition and enteric emissions such as the lactation stage (
Invited review: Large-scale indirect measurements for enteric methane emissions in dairy cattle: A review of proxies and their potential for use in management and breeding decisions.
Furthermore, the environmental performance of PES schemes specifically targeting enteric methane emissions depends on the absence of negative spillovers on other factors of GHG emissions in dairy farms (fertilization management, machinery…). On dairy farms, other important sources of GHG emissions are energy use, manure production and management, fertilizer production and application, and concentrate feed production (
). Using more grass is a lever that can also contribute to lower emissions from concentrate feed and to increase carbon storage in agricultural soils (
). Decreasing enteric methane emissions per liter of milk is a lever particularly interesting to rapidly reduce the contribution of dairy farming to climate change while also considering the food security dimension as it refers to the quantity of food produced. But it should be complementary to further farm-level and area-based assessments of GHG emissions to support effective mitigation strategies.
Impact of increasing the grassland area on the marginal cost of milk production
Several model specifications were tested, and the results are robust to a change (single variable costs equation, system of equations with and without imposing constraints on the parameters for ensuring homogeneity of degree 1). The estimation of variable input shares provides additional information and improves the quality of the variable cost estimation (measured by R2). Consistent with the hypothesis of cost minimization, imposing restrictions on the parameters across equations also improved the variable cost estimation quality. Therefore, we present the results of the constrained system estimation.
The variable cost functions estimated are homogeneous of degree 1 in variable input prices by specification, and we verify they are concave and nondecreasing with input prices (positive estimated variable input cost shares), and monotonic in the milk production level. However, we observe that variable costs increase with the level of fixed factors of production for some observations (see Table A3 in the Appendix). In particular, variable costs are decreasing with more capital for a significant share of farms in 4 models out of 7. Previous studies also found evidence of the violation of theoretical properties on quasi-fixed inputs, in particular capital (
The first-order derivative of the variable cost function (5) gives the marginal cost function (8), in which parameter
corresponds to the effect of grassland surfaces on the marginal cost of milk.
(8)
The results presented in the following paragraphs are calculated from the regression results detailed in the Appendix (Tables A4-A10).
When applied to all dairy farms of the sample and on the subsamples of the western plains production basin and mountainous areas, the model suggests that producing milk with more grass does not significantly affect variable costs (Table 5). There are low but significant extra costs per additional hectare of grassland in plain dairy farms outside the western regions (+0.6€/1000L/ha). This finding illustrates the natural competitive advantage that dairy farms from the western regions or mountainous areas have in producing more milk with grass, the former because of favorable temperate climate (
). A first implication for the design of a PES is that dairy farms from plain areas with a natural disadvantage for producing milk with grass require higher levels of economic incentives in comparison with other production basins. Other authors found evidence of differences in emissions abatement costs among dairy farms according to their geographical location (
Behind the heterogeneous extra costs found in plain production basins also lies a disparity depending on the type of fodder system (Table 5). Considering plain dairy farms, we compare those with a share of corn silage in the fodder area greater than 30%, between 10 and 30% and less than 10%. Extra-costs are significantly negative when the share of corn is higher than 30% (−1.4€/1000L/ha) or lower than 10% (−1.8€/1000L/ha), while they are not significant for an intermediate share between 10 and 30%. This second result reveals the economic opportunity of producing milk with more grass in terms of higher feed self-sufficiency (lower cost feed) (
). For plain grass-based systems with low milk productivity, achieving high self-sufficiency compensates otherwise relatively high marginal costs of production. For “corn silage” plain systems, reducing dependency on high protein content complements represents an important source of cost savings. Current dominant intensive fodder systems involve high expenditures on specific corn inputs (seeds, herbicides, etc.) and high protein content complements (soya, rapeseed) to balance dairy cow feed rations. Synergies between the reduction of GHG emissions and the economic performance of intensive dairy farms have already been pointed out in the literature (
). In contrast, no evidence of this economic opportunity is found for dairy systems with an intermediate share of corn silage in the fodder area. For those fodder systems, higher variable costs from an increase in energy consumption (machinery) and other expenses (seeds, fertilizers, etc.) related to additional pastures and alfalfa management may limit feed cost savings. A second implication of our findings is that currently, a PES for producing milk with more grass with a low level of payment favors adoption by plain fodder systems with a large share of corn silage on the one hand, and fodder systems with a large share of grasslands on the other hand, over intermediate systems combining both corn silage and grass.
Our results suggest that the financial needs for dairy farms to incorporate more grass in their fodder crop rotation are different from one system to another. In particular, we identify the existence and magnitude of extra production costs linked to the modification of fodder crop rotations (increase in grassland area). In the prospect of designing payments for environmental services such as the Eco-Methane program, and aiming to integrate the maximum number of existing farms into the scheme at the lowest cost, we provide evidence of the variability of dairy systems regarding their willingness to accept. With the perspective of accompanying dairy farms to trigger the adoption of mitigation practices at large scale, it is relevant to take into account this heterogeneity, for instance with a differentiated payment or a higher payment for all. Not considering the feeding strategy of dairy farmers, and in particular the type of fodder system, would lower the attractiveness of a payment scheme for some systems.
Given the low level of payment in current PES schemes (for instance in the Eco-Methane program), it seems reasonable to assume that participating farms already had good economic profitability and/or feeding practices compatible with emission reductions when entering the program. In the prospect of engaging more farmers and more liters of milk in the feed ration transition, we identify 3 types of dairy farms that could integrate a PES for the reduction of enteric methane emissions. Farms for which reducing enteric emissions is already profitable (negative extra costs) (type 1), farms for which it requires a financial incentive (no extra costs) (type 2), and farms for which it requires high financial support (positive extra costs) (type 3). Our study suggests that French plain-intensive farms broadly correspond to type 1. Although their individual willingness to accept is likely to be low, the program would still need important financial means to offer a payment given the large number of unit of emissions reductions to compensate (high milk productivity). Grass-based plain dairy systems also belong to type 1 farms, but their lower milk productivity involves lower units of emission reduction to compensate. In contrast, dairy farms located in mountainous areas are more likely to correspond to type 2, and require a payment level high enough to make the adoption of more grass-based diet the profitable option over keeping their current practices. For farms with already a high share of grasslands, increasing milk productivity may be an additional profitable lever to reduce enteric emissions per liter. We also classify plain dairy farms with 10 to 30% of corn silage in the fodder area as type 2 farms involving a large number of emission reduction units to compensate given the large number of potential participants and productivity levels. Finally, we identify plain dairy farms from outside the western region with a natural disadvantage in producing milk with more grass as belonging to type 3 on average. While they represent the most expensive dairy farms to support, they are associated with lower enteric methane emissions per liter of milk according to the Eco-Methane baseline indicator. Higher payment levels would recognize the mitigation efforts they already provide.
However, in the long term, adapting payments to this heterogeneity might not lead to the most optimal allocation to minimize emissions per liter of milk and may introduce a distortion of competition in favor of the most polluting. An optimal scheme should equalise the marginal cost of methane abatement across all farms, for example through a tax on methane emissions. Knowledge of the heterogeneity of abatement costs can then be used to temporarily support the most polluting farms or with a natural disadvantage in improving their performance or in reorienting their activity if they are unable to compete with such a tax.
The positive effect of grass on methane emissions is likely to be partly offset by a drop in the productivity of the cows. Other sources of omega-3 fatty acids such as linseed could also be integrated in the feed ration to maintain productivity (
). Feed complementation is however likely to be a more expensive lever to implement for most farmers, as they are often not produced in farms. Due to the absence of data, the extra-costs of increasing the use of omega-3 sources rich complements could not be estimated. Furthermore, for the Eco-Methane payment to efficiently subsidize the reduction of enteric methane emissions through an increase in grassland areas, the payment will have to cover both the additional costs per liter of milk and the other extra costs per hectare of grass. Beyond impacting production costs per unit of milk, a new hectare of grassland can have a direct effect on farm costs. In this study, we consider only variable costs (intermediate consumption). There may also be fixed costs (specific machinery for grass cultivation, buildings for storage) or other constraints (access to land) increasing the overall extra costs of participation. A study considering all farm costs (variable and fixed costs) found higher GHG emissions abatement costs per liter of milk in large farms (with a high number of dairy cows) compared with smaller farms (
Two aspects of PES design for reducing enteric emissions per liter of milk are particularly crucial to favor its environmental performance. First, the choice of emission indicator should measure farmers' efforts on both cows' productivity and diet, while being easily and regularly implemented on farm at low cost. Second, evaluating farmers' willingness to accept and its variability according to farm type is necessary to define the optimal payment level and scheme's budget ensuring sufficient participation. For a given productivity level, producing milk with more grass have a different impact on milk variable production costs depending on the production basin and fodder system. This research provide more insights into the impact of methane emissions reduction on the production costs of livestock farms, and how to improve support for pressing abatement measures and contribute effectively to achieving climate targets.
ACKNOWLEDGMENTS
This research is funded by the Horizon 2020 programme of the European Union (EU) under Grant Agreement No. 817949 (CONSOLE project, https://console-project.eu/). The authors gratefully acknowledge the Bleu-Blanc-Cœur association for providing reference data on the Eco-Methane programme. We also thank the 2 anonymous referees for their constructive comments that greatly helped improving the revised manuscript. On behalf of all authors, the corresponding author states that there is no conflict of interest.
APPENDIX
Cost function with the translog specification:
Steps to impose homogeneity of degree 1 in input prices:
Parameter constraints to impose to our system:
Homogeneous function to estimate:
Input demand functions: applying Shephard' lemma:
Figure A1Ratio of the Eco-Methane baseline and Tier 2 enteric methane emissions indicators according to milk productivity for different Eco-Methane scenarios. Source: The authors, based on French FADN and BBC data.
Biodiversity protection through results based remuneration of ecological achievement. Report Prepared for the European Commission, DG Environment, Contract No ENV.B.2/ETU/2013/0046.
Institute for European Environmental Policy,
London2014
Démarche Environnementale : La Démarche Éco-Méthane de Bleu-Blanc-Cœur. Environmental Engagement: The Eco-Methane Programme of Bleu-Blanc-Coeur (in French).
Inventaire des émissions de polluants atmosphériques et de gaz à effet de serre en France - Format Secten. National inventory of GHG and air pollutant emissions in France (in French).
Organisation et méthodes des inventaires nationaux des émissions atmosphériques en France OMINEA - 17 ème édition. Inventory methodology report of GHG and air pollutant emissions in France - 17th edition (in French).
Observatoire de l'alimentation des vaches laitières: description des 8 principaux systèmes d'élevage. Edition 2015–2018. Dairy cows' feed observatory: description of the 8 main farming systems. 2015–2018 edition (in French).
Methane output and rumen microbiota in dairy cows in response to long-term supplementation with linseed or rapeseed of grass silage or pasture based diets.
in: Proceedings of the New Zealand Society of Animal Production. volume 71. New Zealand Society of Animal Production,
Invercargill2011: 243-247
in: Fuss M. McFadden D. Production Economics: A Dual Approach to Theory and Applications, Volume I: The Theory of Production. North Holland,
Amsterdam1978: 2-109
in: Stocker T.F. Qin D. Plattner G.-K. Tignor M. Allen S.K. Boschung J. Nauels A. Xia Y. Bex V. Midgley P.M. Climate Change 2013 the Physical Science Basis: Working Group I Contribution to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press,
Cambridge, New York2013: 659-740
Invited review: Large-scale indirect measurements for enteric methane emissions in dairy cattle: A review of proxies and their potential for use in management and breeding decisions.
La Contribution Climat Energie en France : fonctionnement, revenus et exonérations. The Climate-Energy Contribution in France: operation, revenues and exemptions.
Joint Implementation Project FR1000365: “Réduction des émissions de méthane d'origine digestive par l'apport dans l'alimentation des vaches laitières de sources naturelles en Acide Alpha Linolénique (ALA)” (in French).
Weill, P., G. Chesneau, Y. Chilliard, M. Doreau, and C. Martin. 2009. Method to evaluate the quantity of methane produced by a dairy ruminant and method for decreasing and controlling this quantity. Valorex, assignee. Pat. No. WO 2009/156453 A1.
00282-5&id=gr1.jpg){kind=link}
00282-5&id=gr2.jpg){kind=link}
00282-5&id=fx1.jpg){kind=link}