The relative contribution of river network structure and anthropogenic stressors to spatial patterns of genetic diversity in two freshwater fishes: A multiple-stressors approach

International audience1. Recent findings highlighted the central role of the structure of the river network in shaping spatial patterns of genetic diversity in riverscapes. However, the influence of multiple anthropogenic stressors on these patterns may be just as important and the relative impacts of these two types of predictors have rarely been quantified simultaneously in river networks. Here, we contributed to filling this gap by investigating the relative contribution of both network structure and multiple anthropogenic stressors in shaping spatial patterns of genetic diversity in two freshwater fishes (Gobio occitaniae and Phoxinus phoxinus). 2. We focused on two rivers in which the two fish species were sampled along the upstream–downstream gradient. Microsatellite markers were used to quantify genetic diversity from three indices: allelic richness, private allelic richness and genetic uniqueness. Each sampling site was physically characterised according to its position in the network, and was described for multiple anthropogenic stressors including habitat degradation, fragmentation and stocking. This multiple-stressors approach was conducted using a fully explicit and generalisable analytical framework designed to cope with strong collinearity among environmental variables. 3. Overall, the contribution of network structure to the variance in genetic diversity was 1.8 times higher than the contribution of anthropogenic stressors. Both the position of sites along the upstream–downstream gradient and stocking were strong and consistent drivers of genetic variability. Conversely, the local influences of habitat degradation and fragmentation were species-and river-specific, sometimes even varying along the river channel, thus preventing any generalisations. 4. We concluded that the natural structure of networks and stocking strongly influence spatial patterns of genetic diversity in a predictable way, whereas the influence of other human activities may be much more difficult to predict over species and contexts. K E Y W O R D S commonality analysis, freshwater fish, landscape genetics, meta-analysis, stockin