Morphodynamics of a width-variable gravel bed stream: new insights on pool-riffle formation from physical experiments

Field observations, experiments, and numerical simulations suggest that pool-riffles along gravel bed mountain streams develop due to downstream variations of channel width. Where channels narrow, pools are observed, and at locations of widening, riffles occur. Based on previous work, we hypothesize that the bed profile is coupled to downstream width variations through momentum fluxes imparted to the channel surface, which scale with downstream changes of flow velocity. We address this hypothesis with flume experiments understood through scaling theory. Our experiments produce pool-riffle like structures across average Shields stresses t* that are a factor 1.5–2 above the threshold mobility condition of the experimental grain size distribution. Local topographic responses are coupled to channel width changes, which drive flows to accelerate or decelerate on average, for narrowing and widening, respectively. We develop theory which explains the topography-width-velocity coupling as a ratio of two reinforcing timescales. The first timescale captures the time necessary to do work to the channel bed. The second timescale characterizes the relative time magnitude of momentum transfer from the flowing fluid to the channel bed surface. Riffle-like structures develop where the work and momentum timescales are relatively large, and pools form where the two timescales are relatively small. We show that this result helps to explain local channel bed slopes along pool-riffles for five data sets representing experimental, numerical, and natural cases, which span 2 orders of magnitude of reach-averaged slope. Additional model testing is warranted.Peer Reviewe