Sediment transport dynamics near a river inflow in a large alpine lake

Sediment dynamics were investigated in Lake Maggiore, Italy, with field observations from October to mid-December 2012. Three moorings were deployed in Pallanza Bay, a small embayment on the western side of the lake near the Toce River inflow, to measure temperature and currents throughout the water column and suspended sediment concentration (SSC) was estimated with acoustic instrumentation. River intrusions are shown to dominate observed SSC, although a small amount of sediment resuspension was observed at the site of the shallowest mooring during a large wind event that produced strong upwelling of the thermocline followed by downwelling. Although vertical turbulent sediment flux is typically assumed to indicate resuspension and the upward transport of sediment (w’c’ > 0), downward turbulent sediment flux was observed (w’c’ < 0) near the bed during the largest observed intrusion event. The downward turbulent sediment flux significantly contributes to net deposition rates, which are one order of magnitude larger than rates of erosion measured during the two major events observed. Horizontal transport of sediment occurs in vertically confined layers due to buoyancy-driven intrusions. Beneath the intrusions, sediment settles out of the water column at settling rates that appear to be constant with depth based on acoustic Doppler current profiler backscatter measurements. The effective settling velocities needed to produce the observed vertical transport of SSC during an inflow intrusion are one order of magnitude larger than those due to the Stokes settling velocity (ws) alone. Particle flocculation and possible convective instabilities may play a role in generating the large observed effective settling rates. Suspended sediment concentration (SSC) can significantly affect the ecological health and function of lakes and reservoirs. Sediment can regulate primary production by limiting light availability and also by acting as a source of nutrients (Schallenberg and Burns 2004). High sediment concentrations in a lake or reservoir can lead to poor water quality from high turbidity levels as well as decreased basin volume through sedimentation (Morris et al. 2008). The fate of suspended sediments can control the distribution of persistent organic contaminants and can also affect the availability of these pollutants to aquatic species (Schoellhamer et al. 2007). Lakes and reservoirs can act as sinks for many sediment-bound contaminants that can accumulate and deleteriously affect aquatic ecosystems (Mariani et al. 2008). Additionally, clean sediment inputs to the system can potentially lead to natural recovery of certain sediment-bound pollutants. Understanding the mechanisms governing sediment transport and the relative rates of sediment processes such as deposition, erosion, and settling can be key to understanding how a lacustrine ecosystem will be impacted. River inflows are often the dominant source of nutrients, sediments, and contaminants into a lake or reservoir; thus, the fate of these river-borne constituents depends on the hydrodynamics of the river plume as it enters the lake or reservoir (Rueda and MacIntyre 2010). A river flowing into a body of water will lead to an overflow that propagates along the surface if the river water is less dense than the ambient lake water, or, if it is more dense, the river water will plunge beneath the surface as an underflow (Alavian et al. 1992). If the plume reaches a depth of neutral buoyancy, it will intrude into the water column, forming an inflow intrusion. The depth and the vertical extent of the intrusion are dependent on the initial density of the river inflow, the ambient stratification, and the amount of mixing between water masses (Alavian et al. 1992). A laterally bounded or approximately two-dimensional (2D), buoyancy-driven intrusion of thickness h will propagate at its depth of neutral buoyancy with a speed roughly equal to U ! 0:2Nh (Manins 1976; Ford and Johnson 1983), where the average buoyancy frequency across the intrusion is given by N5ðgDq=qohÞ , where q0 5 1000 kg m 23 is the reference density of water and Dq is the density difference across the intrusion. In a large lake, rotation will influence the trajectory of the plume and *Correspondence: [email protected] 1195 LIMNOLOGY and OCEANOGRAPHY Limnol. Oceanogr. 60, 2015, 1195–1211 VC 2015 Association for the Sciences of Limnology and Oceanography doi: 10.1002/lno.10089 can also act to laterally confine the plume against a boundary. In this case, the width of the plume in the lateral direction is given approximately by b $ ðQN=f 2Þ, where Q is the volume flow rate of the intrusion and f is the Coriolis frequency (Imberger and Hamblin 1982). A growing body of research has investigated the complex mixing and hydrodynamics of sediment-laden river inflows in the near field. However, the ultimate fate of suspended sediment from an inflow intrusion in the far field has received limited attention. The vast majority of research investigating the fate of sediments from an intrusion has focused on numerical modeling with limited observations for model validation (DeCesare et al. 2006; Umeda et al. 2006; Chung et al. 2009). This is largely due to the difficulty in measuring sediment concentration as well as the intermittent nature of inflow intrusions. Furthermore, long time records are required to investigate the timing of sediment delivery and the eventual rates of deposition from an intrusion. Observations of sediment laden intrusions have largely been conducted with transects along the length of the plume that provide a snapshot in time (Best et al. 2005; Vidal et al. 2012). A small number of studies have been conducted with moored observations (Chikita 1990; DeCesare et al. 2006). These are limited in spatial extent but allow for investigation of the time dependence of sediment delivery. Although river inflows are often the dominant source of sediment to a lake or reservoir, sediment and its constituents can be reintroduced to the water column through sediment resuspension with subsequent redistribution in a system. Sediment erosion and resuspension are initiated due to shear stress exerted on the bed by the flow. More specifically, resuspension occurs when the bed shear stress, sb, exceeds the critical bed shear stress, scrit, which is assumed to be a property of the bed material and depends on factors including sediment particle size, bed consolidation, and physiochemical properties (Winterwerp and van Kesteren 2004). The bed shear stress can be estimated in field observations with measurements of the near-bed Reynolds stress (Trowbridge et al. 1999). Sediment resuspension has been neglected in prior studies of Lake Maggiore although no direct near-bed measurements have been made (Callieri 1997). The largest bottom shear stresses, and therefore, the most likely locations of sediment resuspension, have been shown to occur where the oscillating thermocline intersects the bottom of the lake (M€ unnich et al. 1992; Gloor et al. 1994; Lorke 2007), which is typically around 20-40 m in Lake Maggiore. Therefore, internal interface motions have the potential to resuspend sediment around the depth of the thermocline and transport it into the lake interior. We show that this mechanism is negligible when compared to sediment delivered by river inflows. The objective of the work presented here is to investigate the dominant mechanisms governing sediment transport in a large alpine lake near a river inflow. The observations are made over a relatively long time period of two and one half months with moored instrumentation measuring velocity and temperature throughout the water column. Adjusted acoustic backscatter measurements are used as a proxy for SSC throughout the water column during the record and near-bed SSC and turbulent sediment flux measurements are obtained with a laboratory calibration (methods discussed in detail below). Near-bed rates of both sediment resuspension and deposition are derived from vertical turbulent sediment flux measurements. Additionally, the moored instrumentation provided observations of horizontal sediment transport rates arising from inflow intrusions between two sites as well as observations of the effective rate of sediment settling through the water column below an intrusion. The content of the article focuses on measurements made during two dominant events, which we refer to as Events I and II. Event I was a wind event that generated thermocline displacement and sediment resuspension. The physics of this event are verified qualitatively with the three-dimensional (3D) hydrodynamic model. Event II was a large inflow intrusion that generated elevated SSC measurements throughout the water column. During the inflow intrusion, the near-bed vertical turbulent sediment flux contributes to net deposition due to the shape of the sediment concentration profile near the bed. We show that river intrusions are the primary source of observed SSC over the deployment and compare the rates of resuspension and deposition during the two events. Additionally, the rate of horizontal sediment transport between two sites is shown to compare well with 2D intrusion velocity scaling and the effective rate of observed sediment settling from an inflow intrusion was an order of magnitude larger than the expected Stokes particle settling.