Dataset for Habitat Puzzles: Linking Primate Invasions and Declines to Landscape Connectivity

Résumé

Dataset for Habitat Puzzles: Linking Primate Invasions and Declines to Landscape Connectivity Abstract The distribution of primate species in Brazil is currently heavily impacted by the exploitation and fragmentation of the Atlantic Forest. This forest is one of the world’s biodiversity hotspots, and its ecosystems are home to an extremely rich fauna. Fragmentation can lead to the decline of certain species, such as Sapajus nigritus (S. nigritus). Conversely, the proximity of roads and built-up areas can lead to the accidental introduction of species with invasive potential that compete with endemic species, including Callithrix jacchus, Callithrix penicillata, and their hybrids. Anticipating and estimating the invasive threat of marmosets and its amplification over time is an important issue for the conservation of declining primate populations. Similarly, knowing which environments are favorable for the spread of capuchins would make it easier to anticipate their dispersal and maintenance. Using data on climate, human density, vegetation cover, and vegetation types, as well as coordinates indicating the presence and absence of these primates, a comparison of generalized linear models and random forest models enabled us to estimate the conditions for their presence. The study took place in the Atlantic forest fragments of the São João river basin in the state of Rio de Janeiro. By extrapolating the probability of the presence of individuals throughout this territory, these models predicted the most suitable habitats. Based on these predictions, and assuming that the resistance to movement of these animals is linked to the suitability of the environment for these species, we were able to model the connectivity between different habitats. Our results show that the probability of marmoset presence increases with decreasing humidity, non-vegetation, and tree cover. The presence of S. nigritus is explained by an increase in the distance to the nearest road. While marmosets now occupy vegetation types under anthropogenic influence, S. nigritus prefers predominantly natural vegetation. Callithrix’s tolerance of anthropized habitats explains its dispersal potential. Comparisons between models showed that random forest models performed better than GLM models. The territory occupied by the species corresponds to a 0.5-fold increase in the area already occupied by marmosets. This dispersal could primarily occur around territories close to their current presence. The possibilities for the dispersal of S. nigritus are quite significant and are not limited to forest fragments, contrary to our hypotheses. We can expect the existing distribution to multiply by 7.5. These results provide important information that will be used by local associations to manage the populations, habitats, and dispersal paths identified. This is all the more important as marmosets compete with the endangered golden lion tamarin.