Determinantes da diversidade e distribuição de mamíferos não-voadores em diferentes escalas espaciais

Abstract

The diversity of non-volant mammals can be affected at different spatial scales by productivity measures, area size, and/or anthropic disturbances. This thesis combined systematic and quantitative reviews on a continental scale with sampling in a tropical forest to investigate how these factors affect patterns of structure in mammalian communities. In the first chapter, the influence of productivity on the richness and abundance of mammal species was quantified on a global scale. Mammalian diversity was positively related to ecosystem productivity for mammalian species of different sizes (body mass), vertical distribution (except arboreal mammals), and feeding guilds (except insectivorous mammals). In general, this result corroborates the idea that mammalian diversity is positively related to increased productivity, regardless of the functional group. Sites with higher productivity are generally associated with more resources and greater ecosystem support capacity, providing greater resilience to human disturbance than less productive sites. Thus, quantifying productivity can help identify important areas for restoration and propose effective guidelines for mammalian conservation. In the second chapter, the magnitude of the relationship between the area of island ecosystems and the diversity of forest mammals in different types of archipelagos around the world was investigated. The results confirmed a positive relationship between area and diversity, with species richness and abundance increasing, in order, on fluvial, oceanic and artificial islands, but not on islands seasonally connected to the mainland. These findings demonstrate that, except for these last types of islands, island area is still a strong predictor of mammal diversity, reaffirming its importance for the conservation of island communities. In the third chapter, the influence of distance to settlements and the number of families per settlement on the probability of occupancy of mammals in the Anavilhanas and Jaú National Parks was evaluated. We also considered the influence of the number of days each camera operated and the body weight of the mammals on the probability of detection. The probability of occupation of the group of species most hunted for food (i.e., deer, peccaries, agoutis, pacas, armadillos) was lower in places closer to human settlements, while the probability of detection of less hunted species and carnivores was negatively related to the body weight of the species. This may suggest that larger, naturally rare, hunted species are less abundant in the system, possibly because they are also the preferred targets of hunters. In the long run, locals will likely need to travel longer distances to hunt equivalent amounts, which incurs a higher cost. Since bushmeat is an important source of protein and livelihoods for local people in the Amazon, illegal and commercial hunting threatens local people's food security. In summary, this thesis emphasises the importance of considering measures of productivity, area, and anthropic factors, in addition to including different functional groups, to access patterns of diversity and distribution of forest mammals at different scales.