Conflicting effects of woody debris on stream fish populations: implications for management

1. Coarse woody debris (CWD) in stream channels causes changes in flow, sedimentation and ratios of pool to riffle areas. There is a consensus among fishery managers and scientists that CWD is beneficial to stream fish communities because of its enhancement of habitat diversity, invertebrate production and cover. Our hypothesis was that CWD accumulation or introduction would not increase in‐stream habitat capacity for all species and their ontogenic stages at reach and stream scales. 2. The study used a system of gravel‐bed streams with naturally dynamic CWD accumulations and a fish community consisting of Salmo trutta, Cotttus gobio, Phoxinus phoxinus, Lampetra cf planeri, Nemacheilus barbatulus and Anguilla anguilla. Cotttus gobio and L. cf planeri are protected by an EU Directive and S. trutta is exploited for angling. Riffles, pools and CWD matrices, considered as the basic habitat/spatial units of channel structure, were sampled separately and abundance of each fish species quantified seasonally at each spatial scale. 3. Multiple‐pass electric fishing techniques were used. Capture efficiencies were calculated for species, habitat and season. Areal densities (number m^?2) were compared for habitat types and season using nonparametric anova . Canonical analysis and stepwise multiple regression were used to show the most influential physical variables on fish density. Densities were also compared by unit volume (numbers m^?3) for pools and CWD matrices to investigate direct three‐dimensional use for cover. Reach‐scale densities for each fish species in relation to habitat composition were made using Spearman rank correlation of habitat‐scale densities with proportionate areas of the different habitat units in the reach. 4. Habitat‐scale densities of bullheads and age 0+ trout were negatively correlated with depth and CWD areas for some seasons. Densities of lampreys, older trout, eels and minnows were positively correlated with depth in some seasons. Water depth had the most consistent influence on fish abundance at the habitat unit scale. Three‐dimensional comparisons of pools and CWD matrices indicated that only trout older than 1+ may use CWD habitats as cover. 5. Reach‐scale densities of 0+ trout and bullheads were significantly correlated with proportion of riffle area and negatively with CWD and combined CWD‐pool habitat area in the reach. Densities of older trout, large eels and lampreys were positively correlated with CWD area and combined CWD‐pool area in some seasons. Inundation of riffles caused by impoundment upstream of CWD accumulations reduced spawning habitat for trout, bullheads, brook lampreys, minnows and stone loach. A trade‐off was an increase in refugia for older trout, minnows and eels. 6. Coarse woody debris accumulation in streams is not beneficial to all species or ontogenic stages in a mixed species population and could severely limit essential habitat areas for some species. Thus, physical manipulation of channels should be implemented only after a thorough study of the habitat relationships of all species present, especially where protected species coexist with target species. The relative importance of in‐stream morphological changes depends on the spatial and temporal scale of the species life histories.