Acoustic turbulence measurements of near-bed suspended sediment dynamics in highly turbid waters of a macrotidal estuary

摘要

Sediment-turbulence interactions near the bed are still poorly understood in highly turbid estuaries, especially in the presence of fluid mud layers. This results primarily from the difficulty in measuring co-located velocity and suspended sediment concentration (SSC) at sufficiently high rate to resolve small turbulent flow scales. In this paper, we show how a set of commercial acoustic and optical backscattering systems known as ADCPs, ADVs and OBSs, can be deployed and used in a complementary way to perform large-scale profilings of tidal current and SSC combined with high-resolution velocity and SSC measurements in the highly turbid near-bed zone. The experiment was done in the Gironde estuary (France) which is well known for its turbidity maximum zone characterized by high SSC values, above 1 g l~(?1) near the surface. A first simple inversion method is proposed to convert the backscattered acoustic intensity measured with ADV into SSC data in the highly turbid near-bed zone. Near-bed SSC data from the OBS are used to compensate for the important acoustic sediment attenuation effect at an acoustic frequency of 6 MHz. No a priori knowledge of acoustic backscattering properties of mud suspensions is required with this calibration procedure. We obtain an attenuation coefficient for mud suspensions of 0.28 m~2/kg at 6 MHz leading to a good agreement between the SSC timeseries from the three ADV receivers and the OBS over the entire tidal cycle. The obtained SSC data are then analyzed with respect to the near-bed velocity, Reynolds shear stress and turbulent kinetic energy (TKE) timeseries in order to identify the relevant sediment transport processes during the tidal cycle. Significant differences in bed shear stress and TKE levels are found between ebb and flood stages with effects on near-bed sediment dynamics. During the ebb, maximum levels of tidal current, bed shear stress and TKE are associated with a reduction of near-bed sediment concentration (from 400 kg m~(?3) down to 100 kg m~(?3)). Bed liquefaction process is assumed to occur at this moment with the presence of highly concentrated mud layer and a possible lutocline at a distance of less than 20 cm above the bed. During the first 1.5 h of flood, turbulent activity remains moderate. The near-bed flood current is then inhibited very abruptly while a sudden increase in SSC occurred above the bed. Assuming that the ADV is able to estimate relevant turbulent erosion fluxes, the co-located velocity and SSC are multiplied and compared with settling flux measurements made onboard under quiescent water conditions. The mean sediment settling fluxes (averaged over 3 min) increase with SSC and are in relative good coherence with fluxes in quiescent water below the hindered regime, for SSC below 15 g l~(?1). Reducing averaging time from 3 min to 30 s allows to increase the range of turbulent fluxes and SSC values, up to 99 g l~(?1). At this scale, fluxes keep increasing quasi-linearly at higher SSC, suggesting the inhibition (delay or reduction) of the hindered settling regime as previously shown by Gratiot et al. (2005) from laboratory experiments. However, the 3-min averaged concentration field remained too low to conclude definitively on the effectiveness of such a process. Further analysis conducted at higher SSC regimes and under fully verified equilibrium are necessary.