Livestock farming generates some striking externalities; whilst it provides 30% of human dietary protein, it occupies 75% of agricultural land, emits 14-17% of anthropogenic greenhouse gas emissions, and uses more antimicrobials than human medicine. Demand for livestock products is high and rising, especially for pork which has quadrupled in the past 50 years. Livestock farming systems vary considerably in the scale of their externalities, but our understanding of how multiple externalities co-vary across contrasting production systems is limited. Research typically focuses on impacts in isolation, and the synergies or tradeoffs among them are assumed. To identify and promote the types of systems that best limit impacts or even carry co-benefits we need to explicitly consider multiple externalities and evaluate them across a wide range of alternative production systems. The main aim of my thesis was to do this for pig production. I recruited, visited, and evaluated over 100 pig farms in the UK and Brazil from those considered to be the most “intensive” through to those certified as Organic. My analyses treat a breed-to-finish system as a datapoint, which may consist of one or several farms (e.g. breeding, rearing and finishing farms). I developed metrics which advanced the quantitative characterisation of farm animal welfare to be compatible with life cycle assessments and to account for both quality of life and the quantity of life years required to produce a unit of product (Chapter 2). I systematically evaluated two externality costs that are commonly perceived to trade off against one another: land use and antimicrobial use (Chapter 3). I found weak evidence of a tradeoff between these externalities but importantly also found several systems characterised by low externality costs in both domains. These systems were spread across different label and husbandry types, and no type was an indicator of systems that performed well in both domains.

I built upon these assessments of one or two costs by systematically contrasting the land use, greenhouse gas emissions, antimicrobial use and animal welfare of as many of my UK and Brazilian pig systems as possible (Chapter 4). I found evidence of positive associations between land use and greenhouse gas emissions, and antimicrobial use and poor animal welfare, but tradeoffs between these pairs of externalities - systems with low land use generally had low greenhouse gas emissions, but high antimicrobial use and poor welfare. Again however, I found systems that carried relatively low externality costs in all domains. I 6 conclude that contrary to prevailing wisdom, tradeoffs among these externalities are not inevitable. In parallel with this detailed work on pig production, I explored the viral zoonotic emerging infectious disease risks of contrasting ways of meeting livestock product demand (Chapter 5). Analyses to date typically ignore how land use affects emerging infectious disease risks. I created a framework that considered risk factors associated with livestock management and land use. I identified significant knowledge gaps and argued these shortfalls in understanding mean we cannot currently determine whether lower- or higher-yielding systems would better limit the risk of future pandemics. My findings challenge many commonly held perceptions about the externalities of farming systems and have important implications for mitigation strategies. My work illustrates the importance of using empirical evidence rather than relying on patchily supported assumptions. I believe that this warrants the systematic testing of other assumed relationships among externalities. I addressed some important knowledge gaps for the pork sector, and more broadly for emerging infectious disease risks, but the same must be done on a much larger scale, spanning other externalities and sectors.