Abstract1. Understanding the mechanisms that drive beta diversity (i.e. b-diversity), an important aspect of regional biodiversity, remains a priority for ecological research. b-diversity and its components can provide insights into the processes generating regional biodiversity patterns. We tested whether environmental filtering or dispersal related processes predominated along the stream watercourse by analysing the responses of taxonomic and functional diatom b-diversity to environmental and spatial factors. 2.We examined the variation in total b-diversity and its components (turnover and nestedness) in benthic diatom species and ecological guilds (motile, planktonic, low-and high profile) with respect to watercourse position (up-, mid-and downstream) in 2,182 sites throughout France. We tested the effects of pure environmental and pure spatial factors on b-diversity with partial Mantel tests.Environmental factors included eight physicochemical variables, while geographical distances between sites were used as spatial factors. We also correlated a and c-diversity, and the degree of nestedness (NODF metric) with environmental variables.

1. Mass effect, allowing species to persist in unfavourable habitats, and dispersal limitation, preventing species from reaching favourable habitats, are the two major dispersal processes. While dispersal limitation can be detected by experimental or modelling techniques, mass effect is more challenging to evaluate, which hampers our ability to disentangle the influence of the environment versus dispersal on species distribution. This is undesirable for biomonitoring programs built on known species-environment relationships. 2. We developed an approach for detection of species influenced by mass effect. We tested it on stream diatoms, a widely used taxonomic group for stream biomonitoring, from four French watersheds. This approach combined (a) an appropriate spatial framework, the asymmetric eigenvector map (AEM), used in species distribution modelling to measure the relative influence of dispersal versus niche processes, (b) an analysis of negative co-occurrence patterns to separate mass effect from dispersal limitation and (c) a measurement of niche breadths to distinguish between non-spatially structured generalists and species influenced by mass effect. 3. We propose that species characterized by low negative co-occurrence values, a high correlation to spatial factors and average to low niche breadths are sensitive to mass effect. 4. Synthesis and applications. We suggest that the sensitivity of species towards mass effect should represent a new ecological trait to be considered for fundamental and applied issues concerning ecology and water quality assessment. Almost all of the species identified here as influenced by mass effect are contributing to the calculation of different diatom-based indices (e.g. Biological Diatom Index or Specific Pollution-sensitivity Index) and should be treated with caution when assigning ecological status classes to water bodies.

en sciences et technologies pour l'environnement et l'agriculture,