Magnetic tracing of fine-sediment over pool-riffle morphology

- We investigate fine sediment dispersion in a gravel bed stream using tracer sand.
- Fines are stored on bars and riffle margins and appear not to be routed through pools.
- Fines did not appear to infiltrate the substrate.
- Tracer dispersion is lumpy and relates to variations in stream energy through pools and riffles.
- Tracer centroid position permitted calculation of tracer virtual velocity.

Field studies documenting fine-sediment (< 2 mm) transport in gravel-bed rivers are rare. For the first time in a fluvial environment, a technique that enhances the magnetic susceptibility of sand is used to trace its longitudinal dispersion and storage. This paper describes the methodology behind the artificial magnetic enhancement of iron-stained sand, and presents the results from sand tracing exercises conducted on two gravel-bed channels with pool-riffle morphology; one unregulated and sinuous in nature (site A), the other regulated and straight (site B), both situated on the River Rede Northumberland, UK. Two tonnes of magnetically enhanced tracer sand was introduced to site A and four tonnes to site B, to provide information on fine-sediment storage dynamics, interaction of fines with the stream bed, and rates of movement, expressed as virtual velocity (Vi). Sand transport pathways appeared to differ between the reaches; for site A, sand storage was found on bars and riffle margins with no storage or signs of transport through pools, and in contrast pool storage of tracer was a key feature shown at site B. Topographic forcing may cause differences in sediment sorting at site A; topographic highs tend to have low sand transport rates with sand grains becoming congested in these areas, whereas topographic lows show higher transport rates resulting in greater dispersion. Supply limitation of sand on the falling limb of the hydrograph may also become an issue in the topographic lows at this site. Hydrograph differences between the regulated and unregulated reaches could also play a role; however this could not be quantified in this study. There was no evidence of sand infiltration into the bed at site A; however marginal evidence for infiltration into the near-surface (0-15 cm) substrate voids was found at site B. The general lack of evidence for significant infiltration may reflect limited availability of void space in substrate framework gravels. Tracer sand was transported over the bed surface, with little vertical interaction with the substrate, despite periods of gravel mobilisation at site A. Vi over the study duration for site A was 2.28 m day− 1, and 0.28 m day− 1 for site B. These values are greater than those calculated using existing predictive equations developed from gravel tracer data, possibly reflecting differences in the mode of transport between bedload and saltation load.