Forecasting Future Food Security through Agent Based Modelling

Abstract

Regardless of what recognition human involvement has played, the consequences of our changing climate will have a negative effect on both agriculture and human well-being. This is expected to be most exacerbated for populations in the developing world where vulnerability and food insecurity already exists. Climate modelling and scenario building has already been an important step in understanding the global impact of climatic change however agent based modelling represents an opportunity for the concept development of food insecurity from a bottom-up perspective. Creating agents from Tanzania’s districts, this research encapsulated the three food security pillars of food availability, access and utilisation; juxtaposing data from the land use and land cover classification (LUCC) data with nationally sourced data regarding human systems. Emerging patterns of food insecurity were recorded and observed by applying linear parameters related to climate change impacts whilst interactions between agents simulated assistance between secure and insecure districts. Using agent based modelling to forecast food security at a district level uncovered insight into the spatial variability of vulnerability within Tanzania. However, similar to most low- and middle-income countries, severe data gaps relating to key factors within food economy and vulnerability context hinder any conclusions drawn. Future endeavours need to focus on downscaling data and predictions to the spatial resolutions required to advise adaptation efforts. If such models are to integrate regional climate models simulating ecological changes or to incorporate further agent interactions at national, sub-national or community level, strengthening the spatial database infrastructure within these countries needs to be addressed in order to accommodate efforts of downscaling.