Variance partitioning of deconstructed tropical diatom communities in reservoirs cascade

Abstract

We used variation partitioning to evaluate significance of local environment, spatial structure and hydrological connectivity in the diatom community variation in phytoplankton and superficial sediments of six reservoirs of southeast Brazil. Common and rare diatom species were represented by different data sets, according to the species relative abundance and frequency of occurrence. To clarify the connectivity effect on the metacommunity organization, analyses were performed with and without a hydrological connectivity matrix as predictor. Results for rare and common species were similar, but explanation power depended on the habitat and the climatic season. The hydrological connectivity predictor proved to play an important role toward explaining the diatom metacommunity dynamics. Consequently, this landscape feature should not be neglected in ecological models of managed rivers.