Using a scaled model to assess the performance of different sediment augmentation strategies in a restored channel

Sediment augmentation is an increasingly popular strategy for restoring rivers, mitigating flooding, and improving fish habitat. However, it is still unclear where along a river sediment seeding produces effective results, or what the fate of sediment is once placed under different flow conditions. Using a set of flume experiments conducted on a scaled pool-riffle reach, we assess the evolution of the planned augmented sediment cover in the Penticton Creek restoration project in British Columbia, Canada. We investigated three sediment seeding patterns described based on the seeding locations through the pool as: Head-seed (HS), tail-seed (TS), and full-seed (FS). For each seed pattern, the reach response to flood events with magnitudes ranging from 2- to 100-year return discharges was assessed. Our results show that while the FS channels are superior at retaining alluvial materials during low floods (i.e., 2-yr), they rapidly lose this ability as the flood magnitude increases. Examining maps of bed erosion reveals that in the FS channels, nearly all the pool area is vulnerable to a high risk of bed scour during high flood events. However, the bed scour only occurred in the HS and TS channels at locations where sediment had been seeded, dispersing eroded materials throughout the pool area. Our findings suggest that for restoring fish habitat in channels with limited sediment supplies, HS and TS seeding patterns are more effective at mitigating the risk of bed erosion during extreme floods. From a practical perspective, an HS or TS needs less sediment to complete than an FS channel, providing a more economic strategy for restoring channels.