An estimation of the sand suspension in alpine rivers during a dam flushing event

Abstract

A significant amount of sand is transported in piedmont gravel-bed rivers that can often be guessed because of the large quantity of sand found in gravel bars. It is however very difficult to evaluate sand fluxes in such rivers because of the concomitant gravel bedload transport and silt and clay suspension (generally assimilated as washload). Also, actual methods to sample sand suspension and surrogate methods such as acoustics remain difficult to apply because of the large velocities and high washload concentrations observed during floods. In this paper, we estimated sand suspension in two gravel-bed alpine rivers during a flushing event (in May 2018 in the Isère River and in June 2019 in the Arc River). Two systems to measure sand flux were evaluated and compared through some simple modelling: a Delft bottle and a peristaltic pump associated to measurements of the flow velocity (ADCP or radar). The objective is to test and validate the pumping system, which is easier to deploy and to collect multiple samples but not as accurate as isokinetic samplers. If the Delft bottle may underestimate the sand flux for high flow velocities, there is a large risk of overestimation using a peristaltic pump since the intake velocity is often much smaller than the flow velocity. Eventually, such experiments allowed us to propose a first evaluation of the sand flux in alpine rivers during a flushing event. It was found that sand fluxes during these events were of the same order of magnitude than Suspended Particulate Matter (SPM) fluxes evaluated from the hydro-sedimentary stations.