Analysis of milk production and greenhouse gas emissions as a function of extreme variations in forage production among French dairy farms.

Many studies that investigate mitigation strategies of greenhouse-gas (GHG) emissions from farming systems often build farm typologies from average data from multiple farms. Results from farm typologies are useful for general purposes but fail to represent variability in farm characteristics due to management practices or climate conditions, particularly when considering consequences of extreme environmental events. This limitation raises the issue of better distinguishing, within datasets of farms, farms that have average characteristics from those that deviate from average trends, in order to improve assessment of how climate variability influences farm performance. We applied the statistical method called Extreme Value Theory (EVT) to identify dairy farms that produced "extreme" amounts of forage. Applying EVT to a dataset of dairy farms from Normandy, Lorraine and Nord-Pas-de-Calais (France) identified subsamples of 10-30% of dairy farms with the smallest or largest amounts of grass from pastures or maize silage in each region. Characteristics of farms with extreme amounts of each forage often differed among regions due to the influence of geography and climate. Farms with the largest amounts of grass or the smallest amounts of maize silage had a variety of cow breeds in Normandy and Lorraine but had only Holstein cows in Nord-Pas-de-Calais. Conversely, most farms with the smallest amounts of grass or the largest amounts of maize silage had Holstein cows, regardless of region. The region also influenced whether farms were oriented more toward producing milk with higher fat and protein contents (Normandy and Lorraine) or toward producing larger amounts of milk (Nord-Pas-de-Calais). As the amount of a given forage changed from smallest to largest, a significant increase or decrease in the amount of milk produced usually changed GHG and enteric methane (CH4) emissions per farm in the same direction as the amount of milk produced. For instance, an extreme increase in the amount of grass fed on farms (1314 vs. 5093 kg/livestock unit/year, respectively) in Normandy was associated with decreased mean milk production (8236 vs. 5834 l/cow/year, respectively) and GHG (7117 vs. 5587 kg CO2 eq./farm/year) and enteric CH4 (3870 vs. 3296 kg CO2 eq./farm/year, respectively) emissions.

Improving Estimates of Nitrogen Emissions for Life Cycle Assessment of Cropping Systems at the Scale of an Agricultural Territory.

In life cycle assessment (LCA), simple models are currently used to estimate cropping system nitrogen (N) emissions on farms. At large spatial scales (e.g., countries), these models are valid. At a smaller spatial scale (e.g., territories), these models may be less accurate, since they completely or partially ignore local conditions such as management practices, soil or climate. The purpose of this study was to consider the variability of those factors when estimating N emissions in LCA at the watershed scale. To this end, Syst'N, decision-support software based on a simulation model of crop and soil N dynamics at field and crop-rotation scales, was applied to predict N emissions from cropping systems in a coastal watershed (Lieue de Grève, France). Syst'N predictions were compared to N emissions estimated by AGRIBALYSE, a static site-dependent method at field and single-crop scales. Syst'N was more sensitive to site-specific soil properties than AGRIBALYSE. Estimates of N emissions that include spatial variability in soil and climate therefore become possible in LCA when a simulation model such as Syst'N is used in the inventory phase.

Environmental impacts of French and Brazilian broiler chicken production scenarios: an LCA approach.

This study compared the environmental burdens of two broiler chicken production systems in Brazil and two in France. One Brazilian system represents large-scale production in the Center-West region of the country; the other is a small-scale production in the South. One of the French systems represents an extensive broiler chicken production system, known as "Label Rouge"; the other is a standard system. Life-cycle impact assessments were performed using the CML-IA characterization method. The main functional unit adopted was 1 tonne of cooled and packaged chicken, ready for distribution. For the systems and impacts studied, production scale did not affect the environmental impact, but production intensity did. The extensive Label Rouge system had the largest impact among the impact categories studied. This resulted principally from the high feed-conversion ratio of this production system (3.1 kg of feed per kg of live weight) in conjunction with the fact that the feed-production stage contributed most to the overall impact. The contribution of deforestation to the crop-production stage was significant, particularly for climate change, equaling 19% of total emissions of CO2eq per tonne of cooled and packaged chicken, in the system of the Center-West of Brazil. The French systems were also affected, since they import crops from Brazil. The system of southern Brazil had less climate change impact because there is no longer deforestation in southern Brazil for crop production.

Emergy evaluation of contrasting dairy systems at multiple levels.

Emergy accounting (EmA) was applied to a range of dairy systems, from low-input smallholder systems in South Mali (SM), to intermediate-input systems in two regions of France, Poitou-Charentes (PC) and Bretagne (BR), to high-input systems on Reunion Island (RI). These systems were studied at three different levels: whole-farm (dairy system and cropping system), dairy-system (dairy herd and forage land), and herd (animals only). Dairy farms in SM used the lowest total emergy at all levels and was the highest user of renewable resources. Despite the low quality of resources consumed (crop residues and natural pasture), efficiency of their use was similar to that of industrialised inputs by intensive systems in RI, PC and BR. In addition, among the systems studied, SM dairy farms lay closest to environmental sustainability, contradicting the usual image of high environmental impact of cattle production in developing countries. EmA also revealed characteristics of the three intensive systems. Systems from RI and PC had lower resource transformation efficiency and higher environmental impacts than those from BR, due mainly to feeding strategies that differed due to differing socio-climatic constraints. Application of EmA at multiple levels revealed the importance of a multi-level analysis. While the whole-farm level assesses the overall contribution of the system to its environment, the dairy-system level is suitable for comparison of multi-product systems. In contrast, the herd level focuses on herd management and bypasses debates about definition of system boundaries by excluding land management. Combining all levels highlights the contribution of livestock to the global agricultural system and identifies inefficiencies and influences of system components on the environment.

An operational method for the evaluation of resource use and environmental impacts of dairy farms by life cycle assessment.

This paper describes and applies EDEN-E, an operational method for the environmental evaluation of dairy farms based on the life cycle assessment (LCA) conceptual framework. EDEN-E requires a modest amount of data readily available on-farm, and thus can be used to assess a large number of farms at a reasonable cost. EDEN-E estimates farm resource use and pollutant emissions mostly at the farm scale, based on-farm-gate balances, amongst others. Resource use and emissions are interpreted in terms of potential impacts: eutrophication, acidification, climate change, terrestrial toxicity, non-renewable energy use and land occupation. The method distinguishes for each total impact a direct component (impacts on the farm site) and an indirect component (impacts associated with production and supply of inputs used). A group of 47 dairy farms (41 conventional and six organic) was evaluated. Expressed per 1000kg of fat-and-protein-corrected milk, total land occupation was significantly larger for organic than for conventional farms, while total impacts for eutrophication, acidification, climate change, terrestrial toxicity, and non-renewable energy use were not significantly different for the two production modes. When expressed per ha of land occupied all total impacts were significantly larger for conventional than organic farms. This study largely confirms previously published findings concerning the effect of production mode on impacts of dairy farms. However, it strikingly reveals that, for the set of farms examined, the contribution of production mode to overall inter-farm variability of impacts was minor relative to inter-farm variability within each of the two production modes examined. The mapping of impact variability through EDEN-E opens promising perspectives to move towards sustainable farming systems by identifying the structural and management characteristics of the farms presenting the lowest impacts.