Taboos and forest governance: informal protection of hot spot dry forest in southern Madagascar.

In the dry forest of southern Madagascar, a region of global conservation priority, formally protected areas are nearly totally absent. We illustrate how the continued existence of unique forest habitats in the Androy region is directly dependent on informal institutions, taboos, regulating human behavior. Qualitative interviews to map and analyze the social mechanisms underlying forest protection have been combined with vegetation analyses of species diversity and composition. Of 188 forest patches, 93% were classified as protected, and in Southern Androy all remaining forest patches larger than 5 ha were protected. Eight different types of forests, with a gradient of social fencing from open access to almost complete entry prohibitions, were identified. Transgressions were well enforced with strong sanctions of significant economic as well as religious importance. Analyses of species diversity between protected and unprotected forests were complicated because of size differences and access restrictions. However, since, for example, in southern Androy >90% of the total remaining forest cover is protected through taboos, these informal institutions represent an important, and presently the only, mechanism for conservation of the highly endemic forest species. We conclude that social aspects, such as local beliefs and legitimate sanctioning systems, need to be analyzed and incorporated along with biodiversity studies for successful conservation.

Patterns of loss and regeneration of tropical dry forest in Madagascar: the social institutional context.

Loss of tropical forests and changes in land-use/land-cover are of growing concern worldwide. Although knowledge exists about the institutional context in which tropical forest loss is embedded, little is known about the role of social institutions in influencing regeneration of tropical forests. In the present study we used Landsat images from southern Madagascar from three different years (1984, 1993 and 2000) and covering 5500 km(2), and made a time-series analysis of three distinct large-scale patterns: 1) loss of forest cover, 2) increased forest cover, and 3) stable forest cover. Institutional characteristics underlying these three patterns were analyzed, testing the hypothesis that forest cover change is a function of strength and enforcement of local social institutions. The results showed a minor decrease of 7% total forest cover in the study area during the whole period 1984-2000, but an overall net increase of 4% during the period 1993-2000. The highest loss of forest cover occurred in a low human population density area with long distances to markets, while a stable forest cover occurred in the area with highest population density and good market access. Analyses of institutions revealed that loss of forest cover occurred mainly in areas characterized by insecure property rights, while areas with well-defined property rights showed either regenerating or stable forest cover. The results thus corroborate our hypothesis. The large-scale spontaneous regeneration dominated by native endemic species appears to be a result of a combination of changes in precipitation, migration and decreased human population and livestock grazing pressure, but under conditions of maintained and well-defined property rights. Our study emphasizes the large capacity of a semi-arid system to spontaneously regenerate, triggered by decreased pressures, but where existing social institutions mitigate other drivers of deforestation and alternative land-use.