Shifting cultivation is one of many phenomena involving the interaction of humans with their natural environment, for which the spatial and temporal dimensions are essential to understanding the system. As a spatiotemporal phenomenon, shifting cultivation uses forest resources to provide the means for human sustenance, both from the temporary use of patches of forest for agricultural production and from the harvest of non-timber forest products from that same mosaic of forest and agricultural land.

There are many factors that affect household decisions to clear forested land for cultivation as well as the permanence of land clearing (deforestation or temporary clearing). Even though subsistence households act independently, their individual decisions have a significant collective impact. Key to addressing the sustainable use of forest resources is a correct understanding of household decision-making and spatiotemporal dynamics. The principal objective of this research is to investigate the spatial and temporal dynamics of forest use by subsistence agricultural households by developing a mathematical model that is able to mimic the spatial patterns (mosaic) of land use typical of subsistence agriculture in southern Cameroon.

More specifically, I develop a conceptual framework for the dynamic choice of spatial location for agricultural production. I investigate household preferences for particular characteristics of forest and fallow land for cultivation and predict them based on household socioeconomic characteristics. The resulting parameter estimates drive a structural simulation model that simultaneously models decisions in space and over time by a heterogeneous set of households representative of the study communities. The model is specified so that it can be adapted to different locations and their specific social and economic contexts. While the ultimate goal is to link the model developed herein to a larger model of livelihood choice at the household level, at its present level of development it provides significant insights into the spatiotemporal dynamics of forest use and shifting agricultural production. It is able to differentiate between the standard Boserup hypothesis of population-induced intensification and model other factors that have a significant impact at the village and household scale. Personal preferences over particular forest or fallow characteristics impact land use intensity. The model demonstrates the importance of path-dependency as a factor in the evolution of the forest-fallow-agriculture mosaic. Finally, I demonstrate its use for comparative dynamic analysis of the impact of exogenous economic shocks and alternate policy scenarios on households and the agriculture-fallow-forest mosaic.