Effects of an amazonian dam on taxonomic, functional and phylogenetic diversity of non-volant small mammals

Hydropower-plant expansion in tropical forests presents environmental and ecological challenges. Our study investigated changes in multiple dimensions of biodiversity of small non-volant mammal assemblages (marsupials and rodents) after the construction of a dam in the Brazilian Amazon, where we expected reductions in number of species and abundance, functional traits, and phylogenetic diversity. We use trap monitoring data for four consecutive years. Our results showed that the number of species and abundance initially increased after reservoir filling but decreased in the second year, reaching levels similar to the Pre-filling period. Species composition changed among the three sampling periods, with higher dissimilarity in the first post-filling year. Functional traits related to diet, locomotion, and body mass were affected by flooding. Frugivorous-omnivorous, terrestrial and larger-bodied species increased in frequency in the first post-filling year, while insectivorous-omnivorous and scansorial species were less frequent in this period. Phylogenetic diversity remained stable throughout the study periods, indicating no significant effect of reservoir filling on this dimension of biodiversity. Our results suggest that dam construction had an immediate impact on small non-volant mammal communities, with changes in species composition and functional traits. However, after two years, non-volant small-mammal assemblages had not fully recovered to Pre-filling levels, indicating ongoing adjustments in response to the new environment. We highlight the need for long-term monitoring programs with sampling before and after the installation of large hydroelectric plants to understand the spatial and temporal impacts on biodiversity. Our study shows the importance of considering taxonomic, functional, and phylogenetic aspects when assessing the ecological effects of large-scale infrastructure projects based on long-term monitoring data.