Taxonomic, functional, and phylogenetic β-diversity patterns of stream fish assemblages in tropical agroecosystems

Abstract

A multi-faceted assessment of diversity is needed to improve our understanding of the mechanisms underlying biodiversity patterns and to reveal the impacts of land use alterations on β-diversity. In this study, we analysed stream fish β-diversity based on taxonomic, functional, and phylogenetic facets in an intensively cultivated tropical region. We sampled 43 stream reaches in the northwest of São Paulo State, south-eastern Brazil. Each sampling site was characterised according to catchment-scale features, landscape dynamic indicators, local-scale features, and distance between stream reaches as network distance (a proxy for dispersal processes). As response variables, we considered taxonomic, functional, and phylogenetic β-diversities coupled with a null-model approach. For each β-diversity metric, we calculated the mean overall value and tested whether the mean value was different from that expected by chance. To examine variation in β-diversity for the three facets and determine the relative contributions of predictor variables, we used a distance-based approach. Taxonomic and functional β-diversities were higher from the expected value under a null model, suggesting that community assembly of these facets was dominated by deterministic processes. In contrast, phylogenetic β-diversity was not different from that expected by chance, suggesting that the lineage composition of these assemblages was random. Furthermore, for all three facets, there was a positive environment-β-diversity relationship that was determined primarily by local-scale features, whereas catchment features and landscape dynamic indicators were not important. In addition, none of the β-diversity facets was correlated with stream network distance, indicating that dispersal processes were not strongly structuring fish assemblages. Our study suggested that although multiple facets of stream fish β-diversity are ruled mainly by deterministic processes (e.g. species sorting), stochasticity is also important in community assembly. An interesting finding was the mismatch between phylogenetic versus taxonomic and functional β-diversity. It is likely that the lack of non-random structure in phylogenetic β-diversity is due to the variation of phylogenetic signal in some functional traits. Given that landscape dynamic indicators were not correlated with measures of β-diversity, we suggest that the recent sugarcane expansion in our study area probably has not critically affected stream fish β-diversity. Also, it is possible that catchment variables presented little variability and did not overwhelm the effect of local environmental variables on β-diversity. In conclusion, our study suggests that even highly disturbed tropical agroecosystems with a pool of species that is probably decimated, can still display a relatively high β-diversity determined mainly by species sorting. These findings suggest key environmental features that must be considered in restoration or conservation of β-diversity in agroecosystems. Specifically, since variation in β-diversity was explained mainly by local-scale environmental gradients, conservation schemes would ideally protect enough sites to capture this entire gradient. Overall, the knowledge of multiple facets can foment more effective conservation and restoration actions by providing a more comprehensive view of the structuring factors of assemblages.