Modelling the impact of rural frontier migration on tropical deforestation.

A major driver of tropical deforestation is rural frontier migration. In this paper an attempt is made to formally describe the human-environment interactions that are manifested in a forested system experiencing a large influx of rural migrants. The Guraferda district in South-West Ethiopia was selected as an exemplary case-study. On the basis of an extensive field surveys in several villages, the relation different social groups in the area were identified: the native population, recent immigrants and investors. For each of these groups their livelihood and their relation with the forest resources was analyzed on the basis of interviews and mapped via remote sensing. To formalize the identified human-environment interactions, an agent-based model was developed. The model simulates the decision-making process concerning deforestation of the identified agent types. The native population consists of shifting cultivators, while the new immigrants are technologically more advanced and are sedentary farmers. For each grid cell of the landscape, utilities for the agent types are calculated. High potential yields increase utility, while proneness to diseases, high population density and the presence of forest decrease utility. Learning behavior is implemented, allowing native agents to learn from migrants and vice versa, thus increasing productivity. Agricultural investors are added to the model as a passive agent type that can grab land from the previous two groups. Results show that immigration started with the forced resettlements in 1985, after which voluntary migrants followed in great number thereby pushing the native population to new frontiers. Ongoing land grabbing by external investors is accelerating this process