| With the 1994 enforcement of the Marojejy and Anjanaharibe-Sud Reserves in the Andapa region of Madagascar, farmers began to experience a closed agricultural frontier. They have subsequently adopted a variety of strategies to increase land productivity under rising demographic pressure. One dominant trajectory of change has been the environmentally-threatening intensification of tavy production (slash and bum), realized through declining fallow cycles. Another has been the environmentally sustainable intensification of irrigated rice and vanilla and coffee production. This study identifies the conditions in which these and other strategies have become prominent. Environmental constraints, institutional weaknesses in water rights and unequally-distributed irrigable land are found to inhibit a broad shift from hill to irrigated-rice production. Farmers are not, however, necessarily increasing output in the tavy system in order to maintain subsistence in rice production. All maintain some tavy, but many are shifting emphasis to the market sector and purchasing rice for consumption. Despite market liberalization in 1994, this study finds that farmers are not choosing the vanilla and coffee trajectory expressly in response to market demand. Rather, most new market production correlates highly with increasing consumption and land pressure. Farmers are responding to the environmental stress caused by declining fallow cycles in their tavy system by increasing market production, and not by shifting to irrigated production. These results are used to assess agricultural change theory, arguing for a greater range of outcomes than such theory commonly promotes.;This study also traces the effects of these strategies to their specific land-cover change and state consequences. Three models are developed to characterize the linkage between production strategies and their landscape outcomes. Monte-Carlo simulation is found to best reflect farmer variability in cover-change and state profiles within single production strategies, and to predict most closely landscape outcomes under different dominant trajectories of agricultural change. Satellite images are used to verify these landscape outcomes. Significant challenges limit the ability to create historical and current land-cover maps from this imagery in agricultural landscapes. These challenges include extremely small parcels, land-cover types with similar spectral properties, and difficulty in gathering historical ground-truth information. |
