Changes in land cover and use and the associated environmental degradation due to human activities have resulted in extreme alterations of tropical ecosystems, especially in headwater streams and their watersheds in the neotropical savanna. Human pressures related to agricultural expansion and urbanization have led to drastic reductions in native vegetation cover, affecting riparian zones and degrading aquatic ecosystem functioning. There is an urgent need to quantify and predict how aquatic communities respond to different intensities of land use to guide conservation and natural resource management efforts. This thesis aims to evaluate how spatial scales influence the relationship between habitat and benthic macroinvertebrate communities and how land use intensity affects the trophic relationships and biological traits of macroinvertebrates. In Chapter 1, we evaluated how the intensity of land use (represented by a gradient moving from native vegetation toward pasture and sugar cane plantations) influences the energy flow and trophic niches of macroinvertebrates. In Chapter 2, we investigated the spatial scales (e.g., catchment, local) that most influence the taxonomic and functional composition of macroinvertebrate assemblages. Finally, in Chapter 3, we examined the impacts of human pressures on the functional diversity of macroinvertebrate assemblages. we showed that the intensity of land use affects benthic macroinvertebrate assemblages, yielding more generalist feeding behaviors with greater overlap of trophic niches (Chapter 1). In addition, environmental variables at the local and catchment scales significantly explained the variations in the taxonomic and functional composition of Ephemeroptera, Plecoptera and Trichoptera assemblages, but land use variables best explained the differences in functional composition among sites (Chapter 2). Finally, we showed that less impacted sites (under reference conditions) had more specialized and more functional diverse macroinvertebrate assemblages compared to disturbed sites (Chapter 3). These results corroborate the idea that biodiversity should be evaluated at multiple spatial scales and that the functional elements of biological communities should be considered when aiming for conservation and the development of predictive tools. This study contributes to a better understanding of the structure and functioning of streams in the neotropical savanna by subsidizing the development of environmental assessment tools. Such approaches will contribute to the development of more appropriate management and conservation measures that will allow for evaluation of the impacts on biota of further degradation of the ecological conditions in tropical streams.