Modelling Regional Grazing Viability in Outback Australia Using Bayesian Livelihood Networks

Abstract

Outback Australia is characterised by multiple competing trajectories to regional social and economic viability, including a tension between agricultural production and other land uses, reflecting broader social and economic values. However, different regions of the outback experience this tension in different ways. In this context, the concept of sustainable livelihoods represents an important way of conceptualising the health and viability of outback regions and the people who live in them. This concept is receiving increased attention in Australia as a way to understand and address the linkages between social and ecological concerns in rural environments. The scope of the research undertaken and reported in this document is to identify and link key social and economic issues affecting the viability and sustainability of livelihoods in Outback regions. Specifically, the research focuses on enhancing our scientific understanding and filling knowledge gaps pertaining to issues of viability and community health in Outback Australia. Our research integrates cultural, social, and economic dimensions with existing ecological and biophysical understanding of these regions. Also improves existing understanding of the network of relationships among livelihood elements that affect natural resource management and regional viability in general. Specific research objectives include: investigating and exploring advanced methodological and modelling techniques such as probabilistic and social networks; linking qualitative with quantitative approaches for social-ecological complex systems; enhancing of the contribution of community-driven decision making on pathways to alternative futures and regional priorities; and understanding regional viability and sustainability of livelihood systems from the “ground-up”. Finally, the researchers’ overarching goal is to assist the Tropical Savannas CRC in fulfilling its unique role in Outback Australia in understanding the multiplicity of environmental, economic, cultural and social dimensions and contributing to the sustainability and management of outback regions. An important issue is understanding the factors that influence sustainable livelihoods in different contexts. These issues are explored through a review of literature on the livelihoods concept in general followed by a detailed case study of the factors affecting grazing livelihoods in the upper Burdekin catchment. The upper Burdekin region is strongly oriented towards pastoralism, with a predominance of owner-operated family-based enterprises. The aim of the upper Burdekin case study is to improve understanding of the factors influencing outback livelihoods through a participatory Bayesian Belief Network approach. In the context of regional outback Australia, this report redefines the concept of sustainable livelihoods as a system of livelihood elements that contribute uniquely, collaboratively and conjunctively to the viability of the region, communities and individuals. In other words, the notion of sustainable livelihoods adopted for the scope of this research, moves away from collective capital accumulation, and represents a more fundamental, generative and emergent mechanism for social, economic and environmental system integration. The report demonstrates the methodological and technical elements of a Bayesian livelihoods network for grazing systems in outback Australia. We define a Bayesian livelihoods network as a probabilistic network of relationships among livelihood elements present in a subjective system of heuristic inference. We use qualitative, participatory and community-driven information to construct a livelihoods network involving issues of viability and sustainability in grazing livelihoods. Beyond the social science basis of our approach we are demonstrating the use of advanced Bayesian network techniques for representing such systems. We describe the model construction process and analyse key drivers of probabilistic elicitation of livelihood elements as graph nodes in the network. We examine different types of nodes and their probabilistic distributions that emerge from (a) self-reported perceptions and inductive inference of citizens and community members; (b) objectively verified elements of the physical and environmental drivers of the livelihoods system, and; (c) heuristically inferred relationships amongst key members and determinants of sustainable grazing livelihoods. We discuss the importance of social science and qualitative research to inform quantitative and inductive paradigms of probabilistic and cognitive inference, using innovative, bottom-up approaches. Finally, the report presents recommendations for future research and the potential role of heuristics in representing dynamic concepts of structure and form in social systems. Fortunately, many outback farming households have a degree of choice in how they go about maintaining a viable livelihood that includes off farm work, diversification into other sectors and financial investments. The drivers affecting these decisions are many and complex. Understanding these is the key to making improved decisions in the long term. The drivers of choices and outcomes presented in a modelling framework must be robust enough to function across a number of theoretical and empirical cases. Such a robust framework renders the use of a heuristic and logical interpretation of actions and outcomes as an essential mechanism of livelihoods representation.