Developing an environmental calculator for application in the beef industry

Abstract

Global greenhouse gas (GHG) emissions from livestock production contribute 18% to total anthropogenic emissions. Emissions from beef and dairy represent three quarters of this figure, and beef production emissions have risen by an estimated 59% in the past half century in response to increasing global population and wealth. In line with international climate commitments, there exists pressure for beef production systems to increase the emissions efficiency of production, and life cycle assessment (LCA) studies have proved a powerful tool to this end. Application of this knowledge to policy is hampered by the heterogeneity of agricultural systems, however, and farm-level GHG accounting tools contribute a flexible, bottom-up solution to this. A variety of such tools are available, but a number of issues hinder their uptake. This thesis therefore set out to a) identify the most important issues affecting the efficacy and uptake of extant farm-level GHG accounting tools and b) develop a farm-level model (AgRE Calc) to address these issues. A review and test of existing tools found that differences in scope and methodology cause substantial differences in results calculated from common input datasets, an issue exacerbated by the methodological opacity of many tools. The empirical test conducted here provides insight into this, and also highlights the need for such tools to maintain simplicity in input data requirements, whilst maximising flexibility and detail in the output. To this end, the impact of cattle ration composition on modelled emissions was identified as a key parameter. The AgRE Calc model was developed to improve this aspect of the methodology, and used to carbon footprint data from a lifetime experiment focusing on beef finishing strategies and diets. Results of this study suggested that high quality grass-based diets have the potential to be as efficient as housed finishes. Additionally, the importance of good-quality, low-granularity activity data to the precision of the footprint was identified, as was the potential for variability in performance within treatments. The study also highlighted the pivotal role of grazing quality in emissions intensity of production. Literature review found that practitioners and models typically broadly estimate this parameter; this approach lacks accuracy and flexibility, so a novel methodology was defined to enable empirical estimation of this variable. Utilising simplistic input data already required by AgRE Calc, a regression model was developed to predict grazed forage digestibility in relation to sward age and nitrogen fertilisation levels. The model predicts decreasing digestibility, resulting in lower performance and higher enteric emissions, as swards age and fertilisation levels decrease. Monte Carlo simulation was also used to provide an estimate of the uncertainty surrounding this variable, and the results suggest that manipulation of pasture digestibility could be a useful mitigation strategy for emissions from extensive beef production. Uncertainty in modelled emissions was a common thread in these studies, and this was explored in more detail. AgRE Calc was developed for a Monte Carlo-based assessment of epistemic uncertainty within farm-level models. The resulting study found that uncertainty in N2O and purchased feed emission factors was the greatest source of farm-level emissions uncertainty. These factors greatly reduce the certainty with which comparisons between intensive and extensive approaches can be made. As such, it is recommended that uncertainty assessment in future form a greater aspect of farm-level and LCA assessments for livestock, and the methods and data compiled as part of this thesis form a basis for accomplishing this through Monte Carlo simulation. Together, these assessments provide a framework for the development of farm-level tools with a view to increasing their usability and relevance. A number of areas in which further progress can be made are identified, and the thesis argues for recognition of the niche filled by farm-level approaches by the developers of GHG accounting methodologies. As such, the thesis as a whole provides a thorough blueprint for advancement of farm-level modelling of GHG emissions, alongside a comprehensive synthesis of the state of the art.