From continental priorities to local conservation: a multi-level analysis for African tortoises.

Abstract

Terrestrial tortoises are the most endangered group of vertebrates but they are still largely ignored for defining global conservation priorities. In this paper, we explored within a hierarchical framework the potential contribution of prioritization studies at the continental scale to the planning of local initiatives for the conservation of African tortoises at the regional level. First, we modeled the distribution of all the African tortoise species, we calculated three indicators of conservation priority (i.e., species richness, conservation value, and complementarity), and we carried out a gap analysis at continental scale. Second, we focused on the most important region for tortoise conservation and performed the same analyses at higher resolution. Finally, we compared the results from the two scales for understanding the degree to which they are complementary. Southern Africa emerged from the continental analysis as the most important region for tortoises. Within this area, the high-resolution analysis pointed out specific core sites for conservation. The relative degree of species protection was assessed similarly at the two different resolutions. Two species appeared particularly vulnerable at both scales. Priority indices calculated at high resolution were correlated to the values calculated for the corresponding cells at low resolution but the congruence was stronger for species richness. Our results suggest to integrate the calculation of conservation value and complementarity into a hierarchical framework driven by species richness. The advantages of large scale planning include its broad perspective on complementarity and the capability to identify regions with greatest conservation potential. In this light, continental analyses allow targeting fine scale studies toward regions with maximum priority. The regional analyses at fine scale allow planning conservation measure at a resolution similar to that required for the practical implementation, reducing the uncertainty associated with low resolution studies.