Cellular automata and GIS based landuse suitability simulation for irrigated agriculture

Cellular automata (CA) is increasingly used to simulate various dynamic courses, e.g. urban spatial growth, forest fire spread and soil desertification. CA simulates these courses with iterative evolvements. Through simple and partial rules to form complex space structures and patterns, CA can express complex systems which are difficult to perform only with mathematical formulas. In combination with GIS Software, such as ArcGIS, it can directly and vividly reflect the geospatial state changes and accomplish many kinds of spatial analysis. This paper introduces CA concept into irrigated agriculture and presents a new CA based GIS (Geographic Information System) approach which incorporates simulation methods and implementation of land suitability evaluation to assess agriculture land use in a GIS environment. In this paper, multi-criteria evaluation (MCE) methodology is utilised to deal with data relating to the environmental conditions in the study area, the Macintyre Brook Catchment, southern Queensland, Australia. The data include soil, topography, landuse, surface and groundwater attributes. Each criterion is processed with a specified method and forms one criterion layer. These criterion layers are regarded as the initial irrigated agriculture conditions in the CA generation courses. They are weighted using an Analytical Hierarchy Procedure (AHP) approach, which is a well-known method of MCE, before being integrated to generate a temporary baseline suitability layer which becomes conditions for the next course of CA generation. The generation courses proceed step by step and each temporary suitability layer is based on the layer which was previously generated. With CA spatial simulation of this kind and essential statistical selection, the resultant suitability layer could be built. All these data processing and modeling procedures are incorporated into a GIS framework. This framework is developed and implemented as a tool, called AHPCA- GIS, using C#.NET computer language. In this tool, MATLAB software embedded components are adopted for computing AHP matrices, CA simulation is performed,in evaluation procedures, and ArcGIS Engine 9.2 development components are applied to complete spatial modeling tasks. The resultant land suitability map of AHP-CA-GIS simulation is derived from the last course of CA generation with a predefined judgment rule of evolvement steady state. It represents land use suitability potential for irrigated agriculture of the catchment, which can help the stake-holders make better landuse planning decisions. The tool is developed with high applicability and flexibility. Some adjustable parameters are open which allows users to rectify the modeling inputs so as to adapt to different scenarios. It increases the scope of application of the GIS simulation and makes it practical for decision-making.