Abstract :
[en] Tropical forest and savanna are globally critical biomes and cover the overwhelming majority of the African continent. They support the lives of nearly a billion people, store a huge amount of carbon above- and below-ground, and host great biodiversity, including the iconic megafauna. The current extent of tropical forests and savannas, however, remains uncertain, because continental studies are based on remote-sensing products that often fail to distinguish tropical forests and savannas from intermediate vegetation states and in patchy landscapes. Here, we developed a very different approach based on the specific floristics of the forest and savanna biomes, using woody species lists from hundreds of sites scattered across Africa. We specifically propose a biome index based on species affinity to the forest or savanna biome that could be used in the future to predict the distribution of tropical forest and savanna following climate change scenario, or in the past to reconstruct past biomes using palynological or anthracological records. We additionally used the largest ever collation of georeferenced herbarium records made available by the RAINBIO project in association with our biome index, to delineate the distribution of tropical forest and savanna across Africa for circa 1400 tree and shrub species. This allowed us to propose a first map of biome specificity, entirely derived from floristic information, and that could be used as a base map for management and conservation, and specifically for the design of re- and af-forestation programs. Biome specificity was also examined across the phylogeny to identify clades associated with specific biome and environmental conditions, and clades transcending biomes. These results, combining information on species composition in specific sites and on species distribution from herbarium records, provided strong insights into the biogeography and evolutionary ecology of the woody flora in tropical Africa.
