Studies attempting to describe fish community structure in shallow riverine environments typically rely on electrofishing and/or visual (snorkelling) surveys, but few have addressed the relative efficiencies of these two methods at estimating fish density and biomass across wide ranges of geography, taxonomy and life history stages. Multiple paired electrofishing and visual surveys were conducted in 18 temperate Canadian rivers in order to obtain community-wide density and biomass estimates from both methods. Partial canonical multivariate analyses were applied to the paired fish community matrices comparing the results of both surveying methods at the taxonomic levels of family, genus and species, as well as size classes within families and species, to assess the particular effectiveness of each sampling method. Although electrofishing estimates of family and species richness were generally greater, snorkelling surveys tended to generate higher density and biomass estimates for different size classes of many salmonid and cyprinid species. Moreover, mean river biomass estimates derived from visual surveying matched those obtained from our best mean river biomass estimates arising from the two methods combined. This study provides empirical evidence that electrofishing and visual survey methods generate different types of information when assessing fish community structure at the family level or by size classes. Our results provide ample background information for determining the most accurate sampling method for a particular fish community assemblage, which is fundamental to fisheries management and research.Rivers are grouped by province, and numbers in brackets refer to the number of sites sampled in each river. Adjusted conductivity is corrected for water temperature. E, electrofishing; V, visual C. J. MACNAUGHTON ET AL.

Alterations to temporal patterns of river flow regimes resulting from damming and flow regulation practices may have negative consequences for freshwater communities. However, little has been performed to develop a holistic approach to assess the effects of hydrologic alterations on fish communities across a wide range of rivers and between different regulation strategies. To address this, we used daily and hourly hydrologic data from gauges in 10 regulated and 14 unregulated Canadian rivers. Building on the Ecological Limits of Hydrologic Alteration concept, hydrologic alterations for many ecologically relevant flow indices were combined to obtain river‐specific hydrologic alteration scores. Extensive community surveys to estimate fish abundance, biomass, diversity indices and habitat guild representation provided data for the derivation of similar river‐specific biotic alteration scores relative to unregulated river conditions. Our results indicate that biological impairment consisting of significant biotic alteration relative to the means from unregulated rivers was directly related to increasing flow alteration scores, with the smallest fish and flow alteration scores observed in run‐of‐river systems and the greatest alteration scores under hydro‐peaking regimes. Our approach not only examined the relationship between river‐specific hydrologic alteration scores and the associated biotic responses, but also provided a more comprehensive assessment of the flow‐response alteration relationship between regulation practices, which may better inform future environmental flow management guidelines. Copyright © 2015 John Wiley & Sons, Ltd.

Summary Hydropower currently accounts for 63% of Canada's total electricity generation and is bound to increase with the energy demands of a growing population. With damming and flow regulation known as major threats to aquatic biodiversity and river and floodplain habitats, an improved understanding of the specific impacts of river regulation is needed for the proper management of these systems. Although interactions among river flow and thermal regimes have been described in the literature, their concurrent influence on fish guild responses has yet to be analysed for temperate rivers. Such an analysis may be used to identify the ecological traits linked with the flow and thermal variables reflecting river regulation. Extensive field surveys were conducted across 25 unregulated and regulated rivers to estimate fish species density and biomass. Fish guild models were developed to characterise morphologic, trophic, reproductive, habitat preferences and behavioural traits, as well as phylogenetic associations. To characterise ecologically relevant components of the flow and thermal regimes of rivers, we calculated indices based on the magnitude, frequency, duration, timing and rate of change in each driver. Model relationships between fish biomass and density estimates were then run using redundancy analyses (RDA) on each type of guild and dominant patterns of flow and thermal variability. Variables representing the magnitude of summer temperatures and intra‐annual flow variability were consistently selected as independent drivers of fish guild responses (>86% of RDA models), clearly showing the importance of integrating thermal regimes in current river hydro‐ecological studies. Fish guild density and biomass were significantly explained (R2Adj = 25–44%) and predicted (R2CV = 35–76%) by flow and thermal variables characterising regimes across unregulated and regulated rivers, whereas total fish density and biomass were not. Fish guild models based on trait–environmental relationships performed better than those based on phylogeny. Our results also showed that the models describing habitat and trophic guilds had the greatest explanatory power (R2Adj = 0.44 and R2Adj = 0.41 respectively). This study identified differences in guild trait–environment relationships across rivers and the guilds most susceptible to changes in flow and temperature conditions resulting from river regulation. In particular, more constant summer temperatures and lower flow variability favoured habitat and trophic guilds over morphologic, reproductive and behavioural guilds. Our results showed that maintaining particular aspects of the flow and thermal regime may be important for ensuring the presence of certain guilds in temperate rivers.

Ecological processes are structured in space and there are important benefits in incorporating spatial information for the analysis of data sets obtained from field studies. Assessing the effect of different flow management practices on river ecosystems is an example where such an exercise is highly relevant. Human activities such as hydroelectric power production are known to modify the temporal variability in river flow. Flow management strategies may have a direct influence on fishes and may trigger complex cascades of interactions involving different features of the river ecosystem. In this study, we performed an assessment of the effect of different flow management practices on fish count density (no. fish/m2), biomass density (g/m2), and species richness. Data were collected in 941 sites located along 28 Canadian rivers. These rivers were either naturally flowing or had altered flows from one of three flow management strategies: run of the river dams, storage with gradual release, or storage with peak release. Each site (300 m2) was surveyed using paired snorkeling and electrofishing techniques; environmental variables (water depth and velocity, and substrate composition) were also measured. The study spanned a broad geographic range (3497 km, geodesic distance) and involved repeated local observations (16–50 sites/river), and was therefore inherently spatially organized. We used spatial modeling to obtain a baseline to estimate the effect of flow management strategies on fishes. Our results indicate that rivers downstream of flow peaking storage dams have, by far, the lowest fish densities (count and biomass) and species richness, whereas those downstream of gradual release storage dams had higher fish biomass density than the unregulated rivers.

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 (*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 R 2 CV \0:66). 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.