Comparison under controlled conditions between multi-frequency ADCPs and LISST-SL for investigating suspended sand in rivers

This paper aims to investigate the possibility and limits of using the laser device LISST-SL to calibrate and validate the multi-frequency acoustic method for assessing the concentration and grain size of suspended sand in rivers by using Acoustic Doppler Current Profilers (ADCPs). Over the last few decades, the methods for investigating suspended sediment using Acoustic Backscattering (ABS) have been gaining increasing acceptance for riverine applications. ABS allows indirect quantification of the suspended sediment using non-intrusive measurements with high temporal and spatial resolution from a moving boat, with a range of several metres. Despite the advancement of these techniques, the use of ABS usually implies calibration and validation in the field, which can be performed by comparing echo recording by ADCP to data from water samples. Sampling and analysis of water samples noticeably increase the campaign cost and limits the reliability of the calibration because of the time and spatial averaging involved in physical sampling. In the present work, the LISST-SL suited to river applications was tested to provide matching data in one point for the validation of ADCP methods in future field work. To this end, laboratory tests were conducted utilising two ADCPs working at 1200 and 600. kHz frequencies and studying the same water volume that was concurrently being analysed by the laser device. Rapid injections of sand (median grain size: 130-250 μm) produced average concentrations of 100-1000. mg/L in a 0.25-m/s-fast and 1.7-m-deep flow. Echo records at the two frequencies were compared to the mean grain size and concentration measured via laser diffraction. Although noticeable differences between the acoustic and laser assessments of concentration were due to (1) inappropriate sampling of LISST-SL at low flow velocity and (2) investigation scarce resolution with respect to actual gradients, the corresponding mean sizes were well correlated. © 2014 Elsevier Ltd.