No safe refuge? Contrasting effects of hunting on rainforest mammal persistence and (re)colonisation

Abstract Defaunation in tropical forests represents one of the most urgent biodiversity crises of our time, driven largely by unsustainable hunting. Understanding how species persist or disappear under intense hunting pressure is critical for designing effective conservation strategies, especially in unprotected areas. We quantified how hunting influences species dynamics in an unprotected Afrotropical rainforest by analysing monitoring data collected from 2016 to 2023 for eleven terrestrial and arboreal mammal species in the Ebo forest, Cameroon. Using Bayesian dynamic occupancy models with correlated detections, we assessed how species occupancy changes along environmental and anthropogenic gradients, explicitly including hunting intensity and hunters' travel cost as predictors for persistence and colonisation. Elevation emerged as the strongest predictor positively influencing the occupancy of all eleven species, with some species exhibiting non‐linear trends along the elevation gradient. Most species generally occupied areas with high travel cost, especially elephant, chimpanzee, crowned monkey and red river hog. Meanwhile, terrain ruggedness had variable effects, with a strong positive effect on chimpanzee and a negative effect on mona monkey. Surprisingly, hunting showed a positive though weak association with the persistence probability of some species, suggesting that the areas used as refuges by wildlife were not exempt from hunting. Species persistence was generally unaffected by travel cost, except for chimpanzees. In parallel, colonisation probability was lower in areas with past hunting history, indicating that previously hunted areas were less likely to be recolonised by wildlife. Similarly, colonisation tended to be higher in areas with high travel cost, suggesting that more accessible areas were less likely for recolonisation. Synthesis and applications . These results imply that hunting not only targets areas used as refuges by wildlife, but also reduces recolonisation potential and may therefore alter source–sink dynamics. We recommend spatially targeted community‐led actions to limit hunting where wildlife still persists through the creation of no‐hunting areas to ensure the long‐term persistence of wildlife within the forest.