A multi-spatiotemporal scale strategy to evaluate factors controlling pebble mobility and its interactions with bedforms in a lowland gravel-bed river

International audienceThis work analysed the factors controling the spatiotemporal variability in bedload mobility in a low-energy gravel-bed meandering river, with a particular focus on interactions between bedload transport distance and bedforms (bar-riffle-pool units). We first addressed the influence of pre-existing morphology, especially bars, on event transport distance, and examined to what extent the maintenance of bedforms requires hydrological events that can move bedload over distances at least equivalent to the length of two successive identical geomorphic units. Specifically, we tracked the displacement of passive integrated transponder (PIT)-tag particles 10 years after their injection, and applied bedload formulae to reconstruct the event transport distance for a 32-year study period. The results showed that event bedload distances for bed surface median size sediments (D50) could not exceed an active channel width, roughly equivalent to 0.3 times the median distance between riffles and between pools. This has two main implications; first, trapping by bars is rare, or perhaps even absent, at the scale of competent events for most particles located at a distance greater than an active channel width from a bar; second, because no event is able to transport bedload over distances at least equal to the length of one riffle-pool unit, maintenance of bedforms must arise from successive minor changes. Furthermore, our data reveal important variability in bedload velocity at the decadal scale according to the year of injection, type of geomorphic unit of injection, and method of injection. This highlights the long-term effects in low-energy rivers of location, timing, and method of injection on velocity estimate. Lastly, we inferred the long-term bedload velocity from PIT-tag data and applied it to bedload formulae. The resulting velocity of 0.951.3 km century-1 is considered an overestimate because the slowdown of tracers seeded at the bed surface due to their progressive mixing in the active layer and their trapping in bars is still ongoing