Rangelands pastoralists and governments: interlinked systems of people and nature.

摘要

We analyse commercially operated rangelands as coupled systems of people and nature. The biophysical components include: (i) the reduction and recovery of potential primary production, reflected as changes in grass production per unit of rainfall; (ii) changes in woody plants dependent on the grazing and fire regimes; and (iii) livestock and wool dynamics influenced by season, condition of the rangeland and numbers of wild and feral animals. The social components include the managers, who vary with regard to a range of cognitive abilities and lifestyle choices, and the regulators who vary in regard to policy goals. We compare agent-based and optimization models of a rangeland system. The agent-based model leads to recognition that policies select for certain management practices by creating a template that governs the trajectories of the behaviour of individuals, learning, and overall system dynamics. Conservative regulations reduce short-term loss in production but also restrict learning. A free-market environment leads to severe degradation but the surviving pastoralists perform well under subsequent variable conditions. The challenge for policy makers is to balance the needs for learning and for preventing excessive degradation. A genetic algorithm model optimizing for net discounted income and based on a population of management solutions (stocking rate, how much to suppress fire, etc.) indicates that robust solutions lead to a loss of about 40% compared with solutions where the sequence of rainfall was known in advance: this is a similar figure to that obtained from the agent-based model. We conclude that, on the basis of Levin's three criteria, rangelands with their livestock and human managers do constitute complex adaptive systems. If this is so, then command-and-control approaches to rangeland policy and management are bound to fail.