1. The roles that streambed geometry, channel morphology, and water velocity play in the retention and subsequent breakdown of leaf litter in small streams were examined by conducting a series of field and laboratory experiments. 2. In the first experiment, conditioned red alder (Alnus rubra Bongard) leaves were released individually in three riffles and three pools in a second-order stream. The transport distance of each leaf was measured. Several channel and streambed variables were measured at each leaf settlement location and compared with a similar number of measurements taken at regular intervals along streambed transects ('reference locations'). Channel features (such as water depth) and substrate variables (including stone height, stone height-to-width ratio, and relative protrusion) were the most important factors in leaf retention. 3. In the second experiment, the role of settlement location and reach type in determining the rate of leaf litter breakdown was examined by placing individual conditioned red alder leaves in exposed and sheltered locations (on the upper and lower edges of the upstream face of streambed stones, respectively) in riffle and pool habitats. After 10 days, percent mass remaining of each leaf was measured. Generally, leaves broke down faster in pools than in riffles. However, the role of exposure in breakdown rate differed between reach types (exposed pool > sheltered pool > sheltered riffle > exposed riffle). 4. In the third experiment, the importance of substrate geometry on leaf litter retention was examined by individually releasing artificial leaves upstream of a series of substrate models of varying shape. Substrates with high-angle upstream faces (were vertical or close to vertical), and that had high aspect ratios (were tall relative to their width), retained leaves more effectively. 5. These results show that streambed morphology is an important factor in leaf litter retention and breakdown. Interactions between substrate and flow characteristics lead to the creation of detrital resource patchiness, and may partition leaf litter inputs between riffles and pools in streams at baseflow conditions.

The near-bed hydraulic environment inhabited by torrential stream fauna was characterized by recording velocity profiles, near-bed velocities, and local wall shear stresses over the upper surface of boulders (or stones) in a mountain stream located in eastern British Columbia. Velocity profiles regularly deviated from a semi-logarithmic shape, and were often found to be "wedge-shaped", due likely to near-bed acceleration resulting from flow constriction over the leading edge of the stones. Local bed shear stress (τ w ) measured using a Preston-static tube (PST) was generally lowest over the leading edge of the boulders, and increased over the upper surface, reaching a maximum near the rear of the stone (upstream of the point of flow separation). Local bed shear stresses measured by the PST were similar to those estimated from the law of the wall (i.e., velocity gradient method based on estimating τ w from the slope of the regression of U on ln(z)) only at locations where velocity profiles were log-normally distributed. Where velocity profiles were wedge-shaped, the law of the wall underestimated τ w compared to the direct measurements using the PST. The possible influences of torrential streambed geometry and relative submergence on near-bed flow parameters and the ecology of stream fauna are discussed.

Local bed shear stress (Τw) is a fluid dynamic parameter of importance in determining the physical and biological characteristics of stream‐bed environments. Unfortunately, it is often difficult to measure Τw under field conditions. Herein we describe the use of a Preston‐static tube, which is essentially a surface‐mounted Pitot‐static tube, to measure Τw in mountain streams. Our results indicate that it is possible to measure local shear stress quickly, consistently, and inexpensively in the field. This technique also provides high spatial resolution, which should allow for detailed in situ studies of local shear stress at scales relevant to lotic organisms. Such information will be invaluable in studies of benthic organisms and hydraulically relevant phenomena in the near‐bottom zone of lotic systems.

Summary 1. After it enters streams, terrestrially derived organic matter (OM) rapidly absorbs water. Using field and laboratory experiments, we examined how this process affected the buoyancy, settling velocity, transport distance and retention locations of four types of organic matter typically found in Pacific coastal streams (‘flexible’ red alder leaves and three ‘stiff’ particle types – Douglas‐fir needles, red cedar fronds and Douglas‐fir branch pieces). 2. Immersion in water rapidly changed the physical characteristics of alder leaves, Douglas‐fir needles and red cedar fronds, which all reached constant still‐water settling velocities after only a few days of soaking. In contrast, the settling velocity of branch pieces continued to increase for 13 days, eventually reaching much higher values than any other OM type. Dried alder leaves became negatively buoyant after only two days of immersion, while other types took substantially longer (up to 24 days) before the specific gravity of all particles was >1. 3. We released saturated OM particles in an experimental channel and found that all particle types travelled further in a fast, shallow ‘riffle’ than a slow, deep ‘pool’. Comparisons with a passive settlement null model indicated that leaves were retained more rapidly than expected in the riffle (by large protruding stones), while the three stiff particle types travelled further than expected (probably due to turbulent suspension) and were retained when they settled in deeper water between larger stones. In pools, passive settlement appeared to dominate the retention of all OM types, with leaves travelling furthest. 4. These retention patterns corresponded well with those observed when saturated OM particles collected in the field were released in two pools and two riffles in a second‐order coastal stream. 5. When the experimental channel and in‐stream data were combined, the retention rates of the three stiff OM types were closely related to calculated Rouse numbers (Rouse number = particle settling velocity/shear velocity), whereas the retention rate of alder leaves was not. This suggests that different physical mechanisms are responsible for the retention of leaves and stiff OM types in shallow streams.

1. Organic matter derived from terrestrial vegetation (detritus) is a key basal resource in many food webs, including those of streams. Once detritus enters a stream, it is either retained or transported, but the rates of within-reach retention depend on structural characteristics of the stream channel, physical properties of the detrital particles and hydrology. Each of these attributes varies in time and space and the magnitude of variation may be affected by structural changes wrought by land-use. We developed a heuristic model framework in an attempt to find a minimal model that captures the essential temporal dynamics of coarse particulate detritus in small streams. 2. The dynamics of our minimal model are driven by several frequently measured variables. Rate of litter breakdown, average transport distance of a particle, discharge and temperature are the primary variables. In addition, we tested the effects of altering assumptions about the nature of particle type and the effect of channel roughness on transport rates. 3. The model produced patterns of detrital standing stocks generally similar to those measured by empirical studies in small streams. Discharge-dependent transport relations are relatively rare, but our assumed pattern of a sigmoid relationship between transport and discharge produced reasonable outputs. 4. We parameterised the model using data for small streams in coastal British Columbia, Canada. The outputs of the model indicate that channel roughness has a more profound effect on small, mobile particles, such as conifer needles, than on angiosperm leaves, such as alder. The model also indicates that the largest proportion of detritus is broken down in situ in all model scenarios. 5. A major limitation to this model and the literature is a scarcity of detail of leaf interception and entrainment processes in streams. Progress towards integrating ecosystem functions and the dynamics of hydromorphology require integrating these physical drivers with biological processes. We present our model to emphasise that progress towards generalisable models of detrital dynamics is possible if links between physical structure and function can be made explicit.