Effects of river scale flow regimes and local scale habitat properties on fish community attributes

Community attributes result from processes operating at various spatial scales. The hierarchic organisation of rivers and the prevailing hypotheses about the variables that affect fish communities make lotic ecosystems particularly suitable to study processes taking place across spatial scales. The general goal of this study was to investigate how river flow regime and local habitat properties explain and predict fish community attributes in unregulated and regulated rivers. Our objectives were: (1) to identify the hydrological indices that best describe the flow regimes of river segments subjected to different regulation types (unregulated, run-of-the-river, storage, and peaking), and; (2) to estimate the relative importance of hydrological indices (river scale), and water depth, water velocity, and substrate composition (local scale) in explaining and predicting local fish community attributes (species richness, total density, and total biomass). We surveyed 880 sites (similar to 300 m(2)) in 25 rivers (14 unregulated and 11 regulated) located in six physiographic regions of Canada. Based on a discriminant function analysis, nine hydrological indices were selected to represent river flow regime. Models of fish community attributes were developed using linear mixed-models (LMM) by nesting sites within rivers and regions. A few hydrological indices along with habitat properties predicted fish community attributes at the site scale (0.43 < cross-validation ). Specific hydrological indices and local habitat properties can be managed to preserve fish community attributes in regulated rivers. Combining variables associated with different scales in a LMM was a powerful and efficient way to model fish community attributes and produced reliable predictive models.