Over half of the phylogenetic diversity accumulated on the African plant tree of life may be eroded under current biodiversity crisis even though threatened species are not evolutionarily unique

Representing ~ 20% of the world’s land, Africa harbours the world’s highest extratropical plant diversity, hosts about a sixth of global plant richness, 8 global biodiversity hotspots and 20% of the world’s rainforest areas. Unfortunately, nearly one-third of Africa's vascular plant species are at risk of extinction and only ~13% of the continent’s land is designated as protected areas (PAs). In the context of global biodiversity crisis, we still have a poor understanding of how the African tree of life might be pruned. Here, we investigated the extent of biodiversity loss in Africa from a phylogenetic perspective. First, we employed a deep neural network model to predict the conservation status of data deficient (DD) and not-evaluated (NE) species (~8500 species with 1 162 937 occurrence points). Then we fitted a negative binomial model to identify hot nodes of threatened species. Finally, we assembled a phylogenetic tree of ~24 000 vascular plants (with known and predicted conservation status) and employed simulations of species loss to reveal the phylogenetic consequences of extinction risk in Africa. Although our best-trained model (validation accuracy: 82.3%; testing accuracy: 67%) predicted that 52.1% of NE and 74% of DD species are not threatened, we found that the ongoing biodiversity crisis would have dire consequences on both taxonomic and phylogenetic diversity (PD). Specifically, we found that 54% of all families are over-threatened (hot nodes of threatened species), i.e., contain more threatened species than would be expected from the fit of the model, thus heightening the risk of losing entire clades. Indeed, we found that if all threatened species go extinction, we would lose 59% of all PD accumulated in the African plant tree of life, leading to a loss of more PD than expected at random, although threatened species are not evolutionarily unique species. Unfortunately, ~72% of threatened species and ~79% of the top-1000-EDGE species are not found in any form of protected areas. Overall, our analysis reveals the extent of biodiversity crisis in Africa and the need for steadfast commitments to i) increased data collection efforts particularly in central African regions and ii) implementation of existing policy for an accelerated recovery.