Interactive multiobjective agro-ecological land use planning: The Bungoma region in Kenya

The development of a third world country requires a conscious balance between different planning and policy issues, such as population growth rate, gross national income, self-reliance and long-term sustainable ecological development. This paper reports on a cross-disciplinary project to design a decision support system (DSS) that aims to assist government policy makers in planning the regional agricultural development of the Bungoma region of Kenya. The DSS is based on the agro-ecological zones (AEZ) model, a previously developed non-interactive optimization model that provides an agro-ecological and economic assessment of various types of land uses, including cash-crops, food production, grazing, forestation and farming. This work extends the decision analytic scope of the AEZ model to explicitly incorporate a multicriteria optimization formulation that facilitates a direct trade-off analysis between the various decision criteria within a user-interactive decision support modeling framework. The DSS uses in-depth information about the Bungoma region, extracted from a large-scale FAO database on Kenya that includes information on various climatic and soil characteristics (e.g., thermal and moisture regime, soil type, slope class) and socio-economic data (e.g., projected growth and product demand patterns) for 90000 agro-ecological cells. At each stage of the analysis our system offers the decision maker several alternative planning strategies with different suggested land uses for over 100 different types of crops, fuel wood and livestock land utilization types for evaluation, allowing the decision maker to take into account trade-offs between a number of planning and policy criteria, including food output, net revenue, gross value of output, self-sufficiency, production costs, arable land use and degree of erosion. Furthermore, the DSS facilitates a direct assessment of the stability of each solution with respect to the food output and erosion in any particular cell and for any given climate scenario. Through a Geographic Information System (GIS) interface, the AEZ system can depict proposed solutions graphically within a PC environment. A PC spreadsheet implementation of our multicriteria DSS is illustrated with real data from the Bungoma district and an expert decision maker.