Evaluation of allocation methods for calculation of carbon footprint of grass-based dairy production.

A major methodological issue for life cycle assessment, commonly used to quantify greenhouse gas emissions from livestock systems, is allocation from multifunctional processes. When a process produces more than one output, the environmental burden has to be assigned between the outputs, such as milk and meat from a dairy cow. In the absence of an objective function for choosing an allocation method, a decision must be made considering a range of factors, one of which is the availability and quality of necessary data. The objective of this study was to evaluate allocation methods to calculate the climate change impact of the economically average (€/ha) dairy farm in Ireland considering both milk and meat outputs, focusing specifically on the pedigree of the available data for each method. The methods were: economic, energy, protein, emergy, mass of liveweight, mass of carcass weight and physical causality. The data quality for each method was expressed using a pedigree score based on reliability of the source, completeness, temporal applicability, geographical alignment and technological appropriateness. Scenario analysis was used to compare the normalised impact per functional unit (FU) from the different allocation methods, between the best and worst third of farms (in economic terms, €/ha) in the national farm survey. For the average farm, the allocation factors for milk ranged from 75% (physical causality) to 89% (mass of carcass weight), which in turn resulted in an impact per FU, from 1.04 to 1.22 kg CO2 -eq/kg (fat and protein corrected milk). Pedigree scores ranged from 6.0 to 17.1 with protein and economic allocation having the best pedigree. It was concluded that when making the choice of allocation method, the quality of the data available (pedigree) should be given greater emphasis during the decision making process because the effect of allocation on the results. A range of allocation methods could be deployed to understand the uncertainty associated with the decision.