Agricultural land use planning : exploring the potential of spatial agent-based modelling (ABM)

This thesis advances capability in agricultural land use planning by considering how human-natural system interactions influence agricultural land use patterns. The research has developed an approach to simulation of different agricultural scenarios through integrating spatial agent-based modelling (ABM) with geographic information systems (GIS). This approach enables the biophysical and the human factors that influence regional development to be integrated into models that can generate a range of future scenarios for planning purposes. The outputs of this research will enable farmers, land use planners and policymakers such as local government to use it as a tool to gain insights into potential agricultural land use changes and new options for irrigated land to assist them in their decision-making. The complexity of land use planning – specifically agricultural land use planning in Australia – raises considerations about what methods and planning support tools might effectively aid decision-making on land and water resources. Too often land use decisions made by farmers and the cumulative impact of those decisions are not reflected in land use planning, nor is their response to change and innovation. This thesis addresses these drawbacks by developing a simulation model designed to explore agricultural land use scenarios through a bottom-up planning approach. The pragmatic nature of this research required an advanced mixed method design to enable qualitative and quantitative data collection and analysis across a multistage process. This study had three stages similar to the Geodesign methodology (notably agent-based design), implemented in its adaptation via Agent Analyst software, the outcome of which was the Crop GIS-ABM. In stage one, a cross section of stakeholders (farmers, planners, local government and State irrigation bodies) were interviewed and surveyed to identify key factors that influence farmers’ land use decisions. These qualitative and quantitative data were merged to frame a conceptual model of farmer decision-making. In stage two, the workflow for the integration of stakeholders’ insights and knowledge with spatial ABM was mapped and the influential decision factors were synthesised into algorithms, and the Crop GIS-ABM was developed. In the third stage, the model was run to simulate a set of agricultural scenarios and explore how farmers make decisions under changing conditions such as introduction of a new irrigation scheme or new alternative crops. The Dorset region in Northern Tasmania was selected as a case study to ‘test’ the simulation model for largely pragmatic reasons of proximity, local relevance, available GIS data as well as experiencing change transitioning from traditional crops to higher value-added production as access to guaranteed water expands. This research demonstrates the potential effectiveness of a spatial ABM as a decision support simulation tool for agricultural land use planning. The Crop GIS-ABM makes an important contribution to the literature on spatial ABM simulation because it includes stakeholders’ insights and integrates both qualitative and quantitative data into spatial ABM. The primary contribution of this research is to enable planning to better capture the cumulative impact of many individual decisions by farmers on land use change. Further research using a similar design methodology could explore the factors driving farmers’ decision-making in other regions of Australia and the world. However, this present research takes urban and regional planning a little closer to a better understanding of how dynamic individual farmer decisions influence the emergence of regional land use patterns and the value of simulation as a planning tool