Assessing the risk of seasonal food insecurity with an expert-based Bayesian Belief Network approach in northern Ghana, West Africa
Food insecurity is still a major global concern. Population growth, poverty, climate variability, and low agricultural productivity, among others, are threatening food provision. Rural areas in West Africa are particularly vulnerable due to low financial and physical capacity and high dependency on agriculture. Programs to support food provision exist, but their effects bear uncertainties under different weather and land use scenarios. Our study focuses on the regional assessment of food provision in the Upper East Region in northern Ghana under land use changes and different weather scenarios by using a Bayesian Belief Network. Especially in the beginning of the rainy season, there is a high risk of food insecurity. Therefore, seasonality needs to be considered in a modeling approach. In addition, we estimated the vitamin A supply indicating the risk of malnutrition. Improving agricultural programs, increasing income by off-farm activities, reducing post-harvest loss, reducing soil erosion, expanding irrigated areas in the dry season and increasing market demand were assessed in order to support food security. The Bayesian Belief Network specifically handles uncertainty and different data types and allows the visualization of complex socio-ecological interactions to communicate to different expert groups, mainly scientists and field officers. A combination of literature, calculations and expert knowledge was required to fill the knowledge gaps in the West African context. About 95 experts were involved in the Bayesian Belief Network development process. An important finding was that an increase in household income, for example by off-farm income, might be better support for people than agricultural programs for providing a contribution to food security under weather uncertainty, provided that income is partly spent on fertilizer. The same is true under increasing population and decreasing total cropland.