Heterogeneidade ambiental e diversidade de peixes de riachos na Amazônia

Amazon streams are highly heterogeneous systems that encompass a remarkable diversity. Due to the increasing threats to these systems, it is necessary to understand how ecological process in natural areas affect streams and their fish biota. This thesis was divided in three chapters and aims to answer the following questions: 1) How much do catchment variables affect the physical habitat of small streams in the Amazon? 2) What is the relative contribution of environmental and spatial variables on taxonomic and functional alpha and beta diversity of stream fish? 3) How much are distinct components of biodiversity (species diversity, taxonomic distinctness, and functional diversity) congruent and how much can they be predicted from catchment variables? Fifty seven streams were sampled across six river basins in the Amazon region. For environmental characterization, a standardized protocol was used to obtain more than 140 local variables, and 11 catchment variables were obtained from aerial images. Fish assemblages were sampled with hand nets during a six-hour period. Stream catchments were divided in two groups based on altitude and slope. These two variables influenced streams habitats, regulating flow velocity and the types and proportions of substrates. The taxonomic and functional patterns of fish assemblages were affected by environmental filters operating at the catchment scale. Nonetheless, variables at the local scale were particularly important to taxonomic and functional alpha diversity. Despite the significant role of environmental filters, limited dispersal was the main driver of variation in fish diversity, indicating a strong biogeographic factor. Finally, various components of diversity exhibited intermediate congruence, which suggests that no single component can describe patterns of fish diversity. In addition, catchment variables alone could not accurately predict diversity patterns, and therefore it is recommended that additional explanatory variables, including descriptors of local ...