In situ high frequency long term measurements of suspended sediment concentration in turbid estuarine system (Seine Estuary, France): Optical turbidity sensors response to suspended sediment characteristics

Abstract The aim of this study is to investigate the complex response of optical turbidity sensors (side- and back-scattering sensors) to Suspended Particle Matter (SPM) characteristics and the consequences when investigating SPM dynamics from long-term high frequency monitoring networks. Our investigation is based on the analysis of a unique dataset of monthly 12 h cycle measurements of SPM characteristics such as turbidity, concentration, floc size distribution, floc density and organic matter content in the macrotidal Seine Estuary (France) between February 2015 and June 2016. Results reveal that despite calibration to a Formazin standard, turbidity sensor response to SPM concentrations (in the range of 7–7000 mg L −1 ) are strongly variable, from the tidal scale to the annual scale and in different compartments of the Seine Estuary. The variability in the calibration relationships is related to changes in the sensor sensitivity according to (i) the sensor intern technology (mainly due to optical geometry) and (ii) the variability in inherent optical properties (IOP) of SPM. Side-scattering optical instruments (measuring scattering at 90°) provide at the annual scale a more stable optical response than backscattering sensor (measuring scattering at angles larger than 100°) for a wide variety of floc size and density in the estuarine environment, while at the tidal scale the backscatter sensors are the most accurate. Sensor sensitivity is strongly affected by floc characteristics, i.e. their median size D 50 , dry density ρ and the scattering efficiency Qb . Results highlight that the median particle diameter contribute to modify the scattering efficiency Qb as well as the dry density: Qb increases with increasing floc size, and for a given floc size, Qb increases with floc density. In this study, the results are next applied to turbidity data from the long-term automated monitoring network in order to estimate SPM concentrations and estimate the related uncertainties.