Different processes explain community biodiversity, considering the relative importance of environmental gradients, biotic interactions or species dispersal capacities. From those mechanisms arise particular beta-diversity patterns, such as turnover and nestedness. If river benthic diatoms are deeply studied, and in particular their relationships with environmental parameters, little is known about their pure spatial distribution patterns. Research still needs to focus on two points: the extent to which those patterns are i) related to the observational scale, and ii) influenced by natural or anthropogenic stress. Here, we propose to tackle these key questions from a large French dataset, gathering 2748 diatom samples collected from 1992 to 2009. In particular, we try to discriminate turnover (Sorensen index) and nestedness (NODF index) patterns from the total community dissimilarity and their relative predominance according to the river Stralher ranks, at different scales (hydroecoregion, biotype, whole hydrosystem). Results are systematically compared to the one obtained with random communities from null models. The potential role of different environmental parameters in the determination of the prevailing pattern is tested with mixed models. Main results show that at each scale of observation, turnover is predominant at low Stralher ranks (in particular in mountainous regions), while nestedness becomes higher downsteam. Natural stressors (mainly related to the altitudinal context) are clearly correlated with turnover. At higher ranks, a loss of specialist taxa is observed, leading to typical nestedness patterns. Anthropogenic parameters, in particular higher orthohosphates concentrations, are significantly related to this pattern. In conclusion, our results tend to support the hypothesis of Gutiérrez-Canova et al. (2013), stating that natural and anthropogenic parameters generate different spatial patterns in community beta-diversity. This may have implications for conservation strategies, which should for example focus on different natural locations when turnover prevails.