FISH ASSEMBLAGES AND STREAM HYDRAULICS: CONSISTENT RELATIONS ACROSS SPATIAL SCALES AND REGIONS N. LAMOUROUX * and F. CATTANE ´ O CEMAGREF, Biologie des Ecosyste `mes Aquatiques, Lyon, France ABSTRACT General relationships between organisms and their habitat, consistent across spatial scales and regions, suggest the existence of repeatable ecological processes and are useful for the management of stream networks. From published data, we defined four guilds of European fish species with contrasting preferences for microhabitat hydraulics within stream reaches. At the scale of stream reaches and across 139 French sites (590 460 fishes sampled), we analysed how fish guild proportions were related to reach hydraulics (proportion of pools vs. riffles %POOL; median discharge by unit width Q50/W). The strongest correlations were observed between two fish guilds and %POOL ( p < 0.001, r 2  0.41) and between one fish guild proportion and Q50/W (p < 0.001, r 2 ¼ 0.10). These reach–scale relationships were consistent across six large French basins, and consistent with the analyses made at the microhabitat scale. Therefore, microhabitat preferences for hydraulics are strong enough to generate consistent reach-scale community responses to hydraulics across regions, despite the influence of other filters such as temperature, nutrient levels or history. The distribution of basic geomorphic features (pools, riffles) in streams and their modification (by dams, weirs and dikes) can modify the proportion of fish guilds by up to 80%, probably contributing to the long- term decline of riffle-dwelling species in Europe. Copyright # 2006 John Wiley & Sons, Ltd. key words: fish guilds; hydraulics; geomorphology; pool; riffle; stream fish community; up-scaling INTRODUCTION Numerous environmental variables affect the structure of stream assemblages at a hierarchy of spatial scales (Poff, 1997; Angermeier and Winston, 1998). It is difficult to quantify the relative impact of each environmental variable. In particular, many physical variables have been related to the structure of communities across stream reaches. These variables include, but are not limited to, discharge rate, slope, stream order, water depth, velocity, shear stress, size of bed elements, type of geomorphologic units and turbulence intensity (e.g. Statzner et al., 1988; Palmer et al., 2000). Most variables also influence communities by their spatial and temporal variations (e.g. Poff et al., 1997; Richter et al., 2003) and are generally inter-correlated across multiple stream reaches. It is, therefore, difficult to identify which physical variables are useful for building simple quantitative tools for the morphological and hydrological management of multiple stream reaches over whole regions. Understanding how and/or if relationships relating organisms to their environment can or cannot be extrapolated across spatial scales contributes to our understanding of the hierarchy of environmental variables affecting stream communities (Levin, 1992; Peckarsky et al., 1997; Jackson et al., 2001; Wellnitz et al., 2001). For example, relationships observed at the scale of individuals (e.g. microhabitat preferences within a stream reach) that are consistent with relationships observed at the scale of communities (e.g. community structure vs. hydraulics relationships across reaches) indicate responses to hydraulics strong enough to persist over spatial scales, despite the influence of other environmental factors (e.g. temperature, nutrients availability) or zoogeography. Consistency across geographic regions also suggests the existence of repeatable mechanisms generating similar patterns in independent landscapes. Recent comparisons of fish relationships with hydraulics across multiple stream reaches suggested that it could be possible to identify key hydraulic variables that are consistently related to characteristics of fish assemblages RIVER RESEARCH AND APPLICATIONS River Res. Applic. 22: 727–737 (2006) Published online 11 April 2006 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/rra.931 *Correspondence to: Dr Nicolas Lamouroux, CEMAGREF, U.R. Biologie des Ecosyste `mes Aquatiques, 3 bis quai Chauveau, CP 220, F-69336 Lyon Cedex 09, France. E-mail: lamouroux@lyon.cemagref.fr Copyright # 2006 John Wiley & Sons, Ltd. Received 10 March 2005 Revised 30 October 2005 Accepted 22 November 2005