Résumé
Human activities’ negative impact on biodiversity is undisputed, but debate remains vivid on their effect on species richness, a key index in ecology and conservation. While some studies suggest species richness declines with human pressure, others show that it can be insensitive or even respond positively to some human pressure, because some species (“losers”) are replaced by others (“winners”). However, many “winners” are favoured by intermediate pressure but decline when pressure becomes too high, and we can thus expect species richness to decline above a certain human pressure. Analysing eBird data in tropical forests, I find that, under a certain threshold, increasing human footprint causes important composition changes with “losers” (habitat specialist, endemic, sensitive and threatened species) being replaced by “winners” (habitat non-specialist, large-range, human-tolerant, anthropophilic and non-native species), resulting in a slight increase in species richness. Above this threshold though, richness in “winners” drops (except for anthropophilic and non-native species), leading to a steep decline in overall species richness. I find that the shape of species richness response to human footprint varies between regions (comparing results from the North America Breeding Bird Survey, PREDICTS database, and eBird data across eight biodiversity hotspots) and identify five different trajectories in species richness response to human pressure. I suggest they can be classified depending on their slope and monotony in the “Replace then remove framework”, unifying contradictory effects of human pressure on species richness.