Agricultural land suitability uncertainty determination using weight sensitivity for climatic inputs in a current and future timeframe

Planners of both public and private infrastructure usually need to take long-term views of how and when land-use will evolve. Agricultural land use is an important determinant of the range and capacity of infrastructure across regional Australia, and climate change is already dictating changing land use in some regions. This paper outlines a process to determine agricultural land suitability, changes of such overtime and the uncertainty in those predictions as a consequence of uncertainty in climatic inputs. Suitability is the evaluation of how well the qualities of a given unit of land match the prerequisites of a specific type of land cover or use. The determination of what areas of land are deemed suitable can be complex and is principally determined through simulated computer models. One commonly used methodology is Multi Criteria Analysis (MCA), which in itself makes use of a variety of modelling techniques. In the modelling of suitability, for any given practice, specific requirements or data-inputs are needed. However, in any modelling techniques uncertainties exist, which can be present within the data inputs or in decisions made in the model construction. Analysis of these uncertainties is rarely undertaken, or applied in any decision making. Model sensitivity analysis can describe uncertainties that are present in the model and the definition of uncertainty boundaries can help in the final definition of suitability outputs. In this paper a method is presented that aims to describe subjective based uncertainties in an Analytical Hierarchy Process (AHP) MCA model for agricultural land suitability across the Gippsland Region of Victoria, Australia. The primary focus will be upon Perennial Ryegrass suitability, although this method has been applied to other commodities across Victoria. Using a one-at-a-time technique on weights within the AHP, uncertainty boundaries based on change points are defined and spatially represented. This is done for climatic inputs for current and projected future timeframes. Through the definition of uncertainty boundaries in suitability modelling, levels of confidence on suitability ratings can be ascribed to final outputs. This can strengthen the capacity of models to describe productive agricultural land. Applications of this could potentially support regional and rural planning decisions into end point land-use, in particular around areas of urban or rural expansion into agricultural land.