Sand and gravel transport through a riffle-pool sequence

Abstract

This study focuses upon flow hydraulics, sediment transport and riffle-pool maintenance on the River Rede, Northumberland, UK. Analysis of bed structure indicate pools to be coarser than riffles, suggesting these to be zones of maximum tractive force at high flow. Tractive force reversal can be demonstrated using a combination of velocity, shear stress and gravel tracer data, and is therefore advocated as a mechanism for maintaining the riffle-pool form. Three dimensional flow structures are likely to increase the likelihood of reversal in pools situated on bends, which may not always be detected using one-dimensional measures of flow hydraulics. Magnetic tracing and basket trapping techniques were used to provide an insight into rates of movement, accumulation, initial motion criteria and routing, of sand. Sand is transported selectively and is mobilised at between 11-22 Nm⁻². Deposition of sub 2mm material is prevalent on morphological high points (bars / riffle margins), although greatest quantities were routed through morphological lows. Freeze core evidence shows limited intragravel storage. Gravel tracer movements showed evidence of size selective entrainment overall, however hiding effects were also found to be evident at two scales; 30-50mm and 110-140mm (for riffles) and 20- 90mm and 11O-140mm for pool. Slope exponents for log-log relations between scaled grain size (D/D₅₀) versus dimensionless shear stress (Өc) of ≈-0.9 suggest that hiding strongly influences sediment transport. Stream power estimates from pgQs demonstrate a higher threshold for motion for gravel in pools (132 Wm⁻²) compared with riffles (127 Wm⁻²). Differences in initial motion criteria (8e) between riffles and pools were found to be significant (p<0.05), indicating pool sediments to be less mobile than riffle, despite pool sediments being less compact. Reduced mobility of pool bedload sediment results from clasts being sheltered by immobile lag gravel found in the pool. It appears therefore that mobility differences between riffles and pools, related to bed structure, does not explain riffle-pool maintenance on the Rede. Scaled travel distance (L/L₅₀s) for tracers in the reach as a whole showed a convex-up relationship with scaled grain size (D/D₅₀s), demonstrating that for tracer grains progressively coarser than the surrounding D₅₀ surface grains, travel distance drops off rapidly, whereas grains progressively finer than the surrounding clasts, travel further but at a less rapid rate. Furthermore, virtual velocity (V*) of tracer grains showed a positive dependence upon D/D₅₀s. Gravel tracer movement provided important insights into riffle-pool maintenance. Transfer of material through the Rede riffle-pool sequence appeared to be influenced by flow magnitude and duration. For low magnitude high frequency flows below 25% bankfull, intra-unit movement was found to predominate. Medium magnitude and frequency flows (up to 50% bankfull) appeared capable of inter-unit transport; scour from pool troughs and deposition on pool exit slopes I riffle heads, movement of material from riffles to bar edges and from bar to bar. For higher magnitude low frequency flows up to bankfull, there was less scour from pools, and a dominance of bar-to-bar sediment transfer. Limited evidence of sediment routing and deposition in pools suggest these to be scour / sediment source zones only, with supply originating from the bed and outer bank. These data demonstrate the importance of different flow magnitude and frequency in creating / maintaining different areas of the riffle-pool structure.