Impacts of global change on a Neotropical poison frog

Global change is causing unprecedented alterations to ecosystems worldwide through multiple drivers, including land use change, pollution, climate change, and emerging infectious diseases. These stressors often interact, producing ecological effects that are difficult to predict when studied in isolation. Such complexity is particularly relevant in tropical regions, which harbour some of the world’s highest biodiversity while also undergoing some of the most rapid anthropogenic transformations. In these systems, amphibians stand out as the most threatened vertebrate group and as especially sensitive to interacting stressors, making them a critical focus for conservation-oriented research. This thesis examines how three major anthropogenic stressors in the Amazon —habitat disturbance, mercury pollution, and infection with the fungal pathogen Batrachochytrium dendrobatidis (Bd)— are associated with variation in habitat use, phenotypic traits, and skin bacterial communities of the poison frog Dendrobates tinctorius. The study was conducted across six populations of D. tinctorius in French Guiana under different gradients of anthropogenic pressure. Data collection combined frog spatial mapping and skin swabbing to assess habitat selection, chytrid infection and microbiome composition, digital photographs to analyse colouration, and water sampling of phytotelmata to quantify mercury concentrations. Across populations, D. tinctorius showed limited variation in habitat selection and no avoidance of polluted phytotelmata. While this apparent stability could suggest persistence under current conditions, it could also indicate reliance on relatively fixed ecological strategies and a dependence on the continued availability of key habitat structures. Notably, the strongest patterns emerged from interactions among stressors rather than from single factors alone. In particular, habitat disturbance seemed to amplify chytrid-related effects on the skin microbiome, whereas the costs of infection alone were not detectable through body condition or colouration. These findings highlight the importance of considering multiple co-existing stressors when assessing species vulnerability and defining conservation priorities for amphibians.