Abstract
The inclusion of key environmental covariables in habitat quality models is crucial to improve the ability of indicators to predict the response of aquatic organisms to anthropogenic stress. The sensitivity of the freshwater blenny (Salaria fluviatilis) to anthropogenic factors was found to be inconsistent among studies, and this might have been because background environmental heterogeneity was not considered. The goal of this study was to assess whether the inclusion of key environmental variables in habitat models would improve the ability to detect the effects of anthropogenic stress on vulnerable fishes. We used a theoretical–information approach to rank different models intended to predict the relative abundance of freshwater blenny in 10 Corsican rivers based on a combination of environmental and anthropogenic indicators. The results show a clear relationship between water velocity and blenny abundance in the studied streams. When this key variable is taken into account in the model, anthropogenic variables such as Enterococcus spp. and Escherichia coli counts, which are indicators of fecal pollution, and the number of competitor species also explain a significant proportion of the variation in blenny abundance. This study reveals the importance of statistically accounting for natural environmental variability, which may conceal other effects, when modeling anthropogenic impacts on a species