Sediment resuspension in Cold Density Currents cascading down lateral slopes of Lake Geneva

Cold-water Density Currents (CDCs) resulting from winter differential cooling and flowing down lateral slopes of a lake (winter cascading) can entrain sediment and contribute to cross-shore transport to its deeper layers. Field investigations along Lake Geneva’s northern shore reveal how CDCs induce sediment resuspension. Acoustic backscattering demonstrated that sediment resuspension is absent at 10-m depth on the steep slope (~30°) just beyond the edge of the shallow coastal shelf where CDCs were initiated but occurred intermittently at 30-m depth where the flow reached higher velocities (slope reduced to ~4.5°). This suggests that CDCs resuspend sediment on the sloping bed (but not on the shelf edge) and potentially transport it to deeper layers, causing sediment focusing (i.e., thicker sedimentation in deeper zones). During CDCs, resuspension occurs in short bursts, often at the head of CDC impulses, creating plumes that can extend to ~1.5 m above the lakebed. Velocity profiles near the bed were well fitted by a logarithmic profile, from which the Shields parameter was determined and compared to the critical Shields stress. However, the strong time-variability of the downslope velocity prevents computation of a representative Shields parameter needed to determine sediment resuspension. The CDC velocity at 1 m above the bed (above ~9 cm s⁻¹) correlated well with high backscattering echo level (indicating sediment resuspension) and was a better predictor for resuspension than the Shields parameter. Since CDCs were found to occur ~25% of the days in winter, CDC-induced resuspension potentially affects lake ecosystem dynamics.

1 Introduction

Sediment resuspension redistributes particulate matter and contaminants over large distances in lakes, significantly impacting aquatic ecosystems (Bloesch, 1995). Sediment resuspension and transport in the nearshore zone of lakes and subsequent sediment focusing, i.e., sediment accumulation in deeper zones is triggered by wind-induced surface waves in shallow areas (Hilton et al., 1986; Rowan et al., 1992; Reardon et al., 2014; Cossu et al., 2017) and internal waves along sloping beds (Shteinman et al., 1997; Hawley, 2004; Boegman and Stastna, 2019). Sediment resuspension and transport in the nearshore zone have also been suggested in flow regimes that are (partially-)driven by differential cooling (Gould and Budinger, 1958; Fer et al., 2002b; Churchill et al., 2004; Peeters and Kipfer, 2009). Valipour et al. (2017) proposed parameterizing resuspension in central lake Erie (Canada/USA) as a function of the instantaneous critical bottom velocity, bottom shear stress and the Shields parameter and to extend the Shields diagram to predict sediment resuspension in lakes.

Cold water Density Currents (CDCs, also called winter cascading) are generated by differential cooling, which occurs when waters on shallow lateral shelves cool faster than in deeper pelagic zones, usually after strong nighttime cooling (Horsch and Stefan, 1988; Peeters and Kipfer, 2009). CDCs, also known as thermal siphons (Monismith et al., 1990), flush the coastal shelf, cascade down the sloping bed until they reach ambient waters with equal density, and then intrude horizontally into the pelagic zone. Such density currents are often highly transient (Fer et al., 2002b; Doda et al., 2023). Earlier studies in Lake Geneva related the dynamics of CDCs to shelf width and cooling intensity which was estimated by the net surface buoyancy flux, $B$ (Fer et al., 2001, 2002a,b). On the lateral slope at 55-m depth, higher acoustic backscatter (indicating greater sediment load) was measured in the upper part of the CDCs than in the water column above (Fer et al., 2002b). However, since no near-bed measurements were available, no correlation between backscattering and sediment resuspension was established in that study. High temporal resolution velocity profiles in the bottom boundary layer are needed to assess resuspension events since they are caused by velocity bursts (Churchill et al., 2004; Valipour et al., 2017). In laboratory studies of density currents flowing down a sloping mobile bed, Maggi et al. (2024) observed that local bursts may increase sediment resuspension or, at least, maintain sediment in suspension, as was found in large eddy simulations (Kyrousi et al., 2018). In oceans, dense shelf water cascading can cause sediment resuspension which can produce substantial suspended and bed load transport toward great depths, in particular, in canyons (e.g., de Durrieu Madron et al., 2013; Palanques and Puig, 2018). This can lead to increased organic carbon transport to the deeper layers (e.g., Palandi de Mendosa et al., 2023). However, detailed measurements, such as reported here, are not available for these large-scale settings.

In this study, we investigate sediment resuspension during winter cascading in Lake Geneva. The following questions are addressed:

• Does sediment resuspension occur in CDCs?

• If so, how intense and sustained is this sediment resuspension and how can it be parameterized?

• Is suspended sediment transported by CDCs from the lateral shallow shelf to the deep pelagic zone?

Below, we analyze a high spatial and temporal resolution dataset of current velocity, acoustic backscattering and temperature profiles of the bottom boundary layer recorded in the sloping nearshore region of Lake Geneva during winter. A detailed analysis of one of these CDC events is complemented by a statistical analysis covering a full winter.

The Supplementary material provides additional figures, tables, and text.

2 Materials and methods

2.1 Field site and measurements

Lake Geneva is a warm, deep oligomictic lake located between France and Switzerland (Figure 1A). The 309-m deep main basin (Grand Lac) is strongly stratified from spring to fall and weakly stratified during winter with a residual thermocline between 100 and 150-m depth (CIPEL, 2021). Occasionally, during exceptionally cold winters, the lake undergoes full-depth mixing, with CDCs playing a fundamental role in cooling/renewing the deepest layers (CIPEL, 2021; Peng et al., 2024). Near-surface water temperatures remain above the temperature of maximum density (except occasionally at the shore) allowing for CDC generation by differential cooling throughout the winter (Peeters and Kipfer, 2009).

Figure 1

Figure 1. (A) Bathymetric map of Lake Geneva with its surrounding mountainous topography (grey shading). Black square: study site. Red triangle: Buchillon Mast, an offshore meteorological station. Red diamond: CIPEL monitoring station (309 m depth), situated in the deepest part of Lake Geneva. Red arrow: mean wind direction during the morning of 13 February 2018. Crosses and annotations outside the map show Swiss coordinates (in meters; CH1903 standard). (B) Close-up of the study site, showing the locations of the three moorings (ME5, ME10, ME30) deployed during winter 2017–2018 (black squares) and where CTD (Conductivity-Temperature-Depth) profiles were taken on 13 February 2018 between 11:30 and 14:30 (local time) starting from the shelf and going towards the pelagic zone (red dots). Contours show isobaths in meters. (C) Data-based schematic of a cold density current (CDC) along a cross-shore transect at the study site, showing temperature profiles taken at locations with depths <40 m (red dots in B), and the position of the three moorings (black squares) with vertical lines (black) of temperature loggers (black ticks). Blue arrows: density current direction. Slanted black line: mean bed slope in the offshore area. Temperature profiles show a CDC with the gray wavy line indicating the upper limit of the density current. Bathymetric data: Géodonnées Etat de Vaud (2016a,b).

Measurements were carried out on the northern shore of Lake Geneva from 17 December 2017 to 22 March 2018 at a site (Figures 1A,B) where winter cascading was previously reported (Fer et al., 2001, 2002a,b). At this site, there is a 250–300 m wide shallow (2–6 m deep) coastal shelf. A wider shelf to the east can provide additional cold water during cooling periods (Fer et al., 2002b; Figure 1B). Perpendicular to the shore, the shallow shelf is followed by a steep section between 6 and 20-m depth with local bed slopes up to 30° (Figure 1C). Further from shore, the bed slope is ~4.5°. Comparable nearshore areas exist at several locations along the northern shore of the lake.

Three moorings perpendicular to the shoreline at 5-m (ME5), 10-m (ME10) and 30-m depth (ME30) had vertical lines of temperature sensors (31 sensors in total; accuracy ±0.002 °C; time resolution 2–5 s; vertical resolution 0.6–3 m; Figures 1B,C) with the lowest sensors at 1, 0.6 and 1 m above bed (mab), respectively. Mooring data, locations and the ME label were used by Reiss et al. (2020) to study coastal upwelling. Upward-looking Acoustic Doppler Current Profilers (ADCPs) were placed on the bed at ME10 and ME30. At ME30, a 2-MHz Nortek Aquadopp-HR Profiler collected high-resolution data in the bottom boundary layer between 0.4 and 3 mab (bin size: 15 cm; time resolution: 5 min). A second ADCP at ME30, a 300-kHz Teledyne RDI Workhorse Sentinel resolved currents between 4.6 and 26.6 mab (bin size: 2 m; time resolution: 20 min). At ME10, a 1-MHz Nortek Signature recorded currents between 1 and 8.8 mab (bin size: 0.6 m; time resolution: 10 min). Corrected echo levels from the Aquadopp-HR and the Signature were determined using the conversion of the raw acoustic backscatter signal and a range normalization, following the Nortek procedure (Lohrmann, 2001; Supplementary Text 1). Profiles from the two ADCPs at ME30 were combined and smoothed with a temporal resolution of 20 min to obtain continuous, full-depth mean velocity profiles. This allowed determining the height of the CDC. The Aquadopp-HR data were used at their full spatial and temporal resolution to study the bottom boundary layer dynamics. Currents were recorded in a local coordinate system, with $U$ being the absolute velocity of the current tangent to the bed, and alongshore $U_{\mathit{AL}}$ and cross-shore $U_{\mathit{CS}}$ components. Measurement accuracy of $U$ was ~0.5 cm s⁻¹. Details of the instruments and mooring settings are summarized in Supplementary Table 1.

CTD (Conductivity-Temperature-Depth) profiles were collected after the cold night of 12–13 February 2018 with low wind (mean wind speed of 1.87 m s⁻¹) to obtain a detailed temperature field along the cross-shore transect and to determine whether the CDCs reached the deep, weakly stratified hypolimnion (see Figures 1B,C, Supplementary Figure 1 for the CTD locations). Air temperature, pressure, wind, and relative humidity data were collected at Buchillon Mast located on the Lake Geneva shelf (~75 m offshore), ~2.7 km to the west of the field site (Figures 1A,B). Net surface buoyancy flux, $B$ , was computed to determine periods of lake surface cooling, when $B>0$ (Fer et al., 2002b; Doda et al., 2022; calculation details in Supplementary Text 2).

For the estimation of the critical threshold for resuspension, surface sediment samples were collected in 2021 near ME30. The sediment grain size distribution was determined using a Beckman Coulter LS 13320 Particle Size Analyzer for four sediment subsamples. The grain size analysis was repeated on three other subsamples following deflocculation for 1 h in a Sonifier with Na₄P₂O₇ solution (Mikkelsen and Pejrup, 2000). The median grain sizes for both cases, i.e., before and after deflocculation (7 sub-samples in total), were computed, and a representative grain size value $d=23\mathit{\mu m}$was determined as the average of both cases. Choosing any value within the measured range does not change the interpretations and conclusions of the study since the threshold of sediment entrainment is not sensitive to $d$ in this range (Soulsby and Whitehouse, 1997; grain size distributions in Supplementary Figure 4). This also supports the use of a single threshold for general resuspension, independent of the sediment size classes, as is commonly done (Graham et al., 2016; Valipour et al., 2017). In Lake Geneva, surface sediment has low cohesiveness and low organic matter content (Loizeau et al., 1997, 2012).

2.2 Cold Density Currents (CDCs): data analysis

2.2.1 Detection of CDCs and sediment resuspension

CDCs were detected based on water temperature and velocity measurements taken at ME10 and ME30 following the procedure described in Supplementary Text 3. Conditions were carefully implemented to select cross-shore downslope flows triggered only by differential cooling. We focus first on a well-defined cascading event, referred to as $E$ , that occurred on 13 February 2018. Potential resuspension was estimated by calculating the Shields parameter $\theta$ (dimensionless bed shear stress) from velocity profiles (Figure 2), its critical value for resuspension, ${\theta }_{\mathit{\beta cr}}$ (Supplementary Text 4), and by examining ADCP backscattering levels. The event $E$ results were then generalized with an analysis of the full-winter dataset.

Figure 2

Figure 2. At mooring ME30 during event $E$ (Figure 3): (A) vent-averaged profiles (thick colored lines) of temperature (blue), and of cross-shore velocity ( $U_{\mathit{CS}}$ ; red; positive downslope) and absolute velocity ( $U$ ; black). Thin lines: 20-min-averaged profiles (sampling frequency of the RDI ADCP). (B) Close-up profiles of absolute velocity $U$ near the bottom boundary (from the Aquadopp-HR); mean profile in red. Thin black lines: selected individual profiles (sampling frequency 5 min). Dashed lines: fitted logarithmic profiles (Equation 1).

2.2.2 Logarithmic profile

Bed shear stress, ${\tau }_b={\rho }_w{u}_{\ast }^2$ , needed to compute $\theta$ , was calculated using the von Kármán-Prandtl Law-of-the-Wall to determine friction velocity $u_{\ast }$ ( ${\rho }_w$ is the water density). Assuming that close to the bottom boundary, the flow is well-mixed due to high shear, the effect of stratification on the velocity profile (Turner, 1973) can be neglected (Kneller et al., 1999; Valipour et al., 2015). The lower part of individual Aquadopp-HR velocity profiles was fitted to the logarithmic profile (Figure 2B; Supplementary Text 5):

$\begin{array}{ll}u\left(z\right)=\frac{u_{\ast }}{\kappa }ln\left(\frac{z}{z_0}\right),& (1)\end{array}$

where $z_0$ is the effective bottom roughness, $\kappa$ ≈ 0.41 is the von Kármán constant, and $u\left(z\right)$ the current velocity at height $z$ above the bed.

3 Results

During the night of 12–13 February 2018, strong and steady differential cooling developed due to very cold meteorological conditions and low winds, resulting in $B>0$ ( $B$ positive for net cooling of the lake surface; Figure 3A, Supplementary Figure 1). A CDC was observed during the morning of 13 February (event $E$ ) with pronounced downslope velocities at the bottom of the water column, associated with cold water flowing underneath warmer ambient water (Figures 1C, 3B,D; Supplementary Figures 1, 2). Velocity and temperature profiles during event $E$ , from 7:00 to 10:00 (local time; Figure 2) showed that a ~0.4 °C temperature drop in the bottom boundary layer matched a rapid increase of cross-shore velocity at 1 mab, $U_{\mathit{CS}1}$ (Figures 3B,D; Supplementary Figure 2). During the period shown in Figure 3, water temperature at ME5 (not shown) was always lower than at ME10 and ME30, confirming that the shelf is the source of cold water. Throughout the CDC, large temperature fluctuations at short time scales were observed, with changes up to 0.2 °C in 30 s at ME30 (Supplementary Figure 3).

Figure 3

Figure 3. Temporal evolution on 13 February 2018 at ME30. (A) Hourly net surface buoyancy flux ( $B$ , solid blue line; blue dashed line: zero; positive flux: heat loss), mean wind speed (red line) and wind direction (red arrows; North up). (B) Temperature from 1 to 18 mab. (C) Aquadopp-HR echo level (0.4 to 3 mab). (D) Shields parameter ( $\theta$ , green dots), its critical value ( ${\theta }_{\mathit{\beta cr}}$ , purple dashed line), absolute ( $U_1$ , blue dashed line) and cross-shore ( $U_{\mathit{CS}1}$ , solid blue line, positive is offshore) velocities, at 1 mab. Blue shaded areas show periods of CDCs; CDC event $E$ lasts from 7:00 to 10:00 (blue double arrow). Red dashed line: time when high $U_1$ and $U_{\mathit{CS}1}$ velocities coincide with high $\theta$ values, and spikes of high echo level and cold temperature.

3.1 Velocity profiles

The velocity profile shape of $U$ and $U_{\mathit{CS}}$ during event $E$ at ME10 (Supplementary Figure 5B) and ME30 (Figure 2) is typical of a density current. The maximum velocity occurs close to the bottom boundary in the cold (dense) layer and the velocity decreases linearly above the height of maximum velocity in the mixing layer as reported by Thorpe and Ozen (2007). Individual profiles show strong temporal variability in strength and profile shape (Figure 2), as previously observed by Doda et al. (2023) in a small, wind-sheltered lake (Rotsee, Switzerland). Such bursts were also found in dense shelf water cascading in the ocean (e.g., Palandi de Mendosa et al., 2023). The strongest CDC velocity burst reached 10 cm s⁻¹ at ME30 at ~1.5 mab (Figure 2B) and temperature profiles indicate that the CDC extended up to height h ≈ 12 mab (Figure 2A). The ratio of the height of maximum velocity ( $h_{\max }$ ) over the CDC thickness was $h_{\max }/h\approx 0.11$ , as previously observed for similar events (Fer et al., 2002b). At ME30, both individual near-bed $U$ profiles and their mean overall matched a logarithmic profile (Equation 1; Figure 2B; Supplementary Text 5).

3.2 ADCP Echo level

Impulses of high Aquadopp-HR echo levels, interpreted as sediment bursts, were observed in the lowest bins during the CDC, in particular, at its head (Figure 3C). Most often, these impulses did not exceed 1.50 mab in height and high echo level values were confined to the near-bed layer (Figure 2A) where strong stratification reduces vertical diffusion, and thus creates stable layers that limit vertical exchange of sediment with the layers above. This can explain the sharp contours of resuspension bursts and suggests that layered flows may travel further.

The background echo level within the CDC, i.e., the echo level without considering near-bed impulses, was lower than in the ambient water before and after the passage of the CDC. This is also seen during a shorter CDC at 2:30 and a weaker CDC between 11:00 and 15:30 (blue shaded periods in Figure 3D). Such lower background echo levels in the CDCs were also observed at ME10 where velocities are lower than at ME30 (Supplementary Figure 5), suggesting that it is a characteristic of the cold water with a low particle concentration being flushed from the shelf. This explains the echo level increase between 10:00 and 11:00 when the CDC briefly stopped.

3.3 Shear stress and sediment resuspension

Critical shear stress ( ${\theta }_{\mathit{\beta cr}}$ ) and Shields parameter ( $\theta$ ) are shown in Figure 3D. It was found that $\theta$ was highly variable during $E$ and intermittently exceeded ${\theta }_{\mathit{\beta cr}}$ , indicating that resuspension can theoretically be expected (Soulsby and Whitehouse, 1997; Graham et al., 2016; Valipour et al., 2017). However, the strong temporal changes of the velocity profile shape led to different estimates of $\theta$ for similar maximum velocities (Figure 3D). Note that $\theta =0.054$ computed from the mean velocity profile is lower than the ensemble-averaged value of $\theta$ from the individual profiles (0.070), and much lower than ${\theta }_{\mathit{\beta cr}}$ ; this is due to short bursts with high $\theta$ values.

Throughout event $E$ , periods during which $\theta$ exceeds ${\theta }_{\mathit{\beta cr}}$ , generally coincided with periods when high acoustic echo level impulses were present. However, at the time scale of individual profiles (sampling time 5 min), $\theta$ values above the threshold for resuspension did not always correspond to the high echo levels (Figures 3C,D), which suggests that the use of $\theta$ to determine sediment resuspension in CDCs is questionable.

3.4 Whole winter statistics

Comparable results concerning the velocity profile shape and the echo level were found during other measured CDC events (examples in Supplementary Figures 6–8). Notably, there were more strong echo level impulses at the head of CDCs and a lower background echo level within the CDCs than in the ambient water. CDCs occurred on 23 days (~25% of the days) during the measurement period from 19 December 2017 to 22 March 2018. For all those CDCs, the echo level close to the bed (in the second Aquadopp-HR bin at 0.55 mab) is plotted against the Shields parameter $\theta$ (Figure 4A) and the velocity at 1 mab, $U_1$ (Figure 4B). Given the large scatter in each panel, binned averages were calculated. Above a threshold of $U_1$ ≈ 0.09 m s⁻¹, velocity is correlated with echo levels (R² = 0.71, Figure 4B), as was seen during $E$ (Figure 3). The correlation was lower using $U_{\max }$ . In contrast, a low correlation was found between the echo level and $\theta$ , even above the theoretical threshold ${\theta }_{\mathit{\beta cr}}$ (Figure 4A). Varying ${\rho }_s$ and $d$ did not modify the correlation strength and the significance of the relation shown in Figure 4A; only the slope of the linear regression changes. These results further question the use of $\theta$ to determine sediment resuspension and transport in CDCs with strong time variability (see Discussion). On many dates when $U_1$ ranged from 5 to 9 cm s⁻¹, echo levels were relatively high. This confirms that intermittent resuspension impulses also occurred during other events in the dataset with current intensity comparable to event $E$ .

Figure 4

Figure 4. Echo level from the second Aquadopp-HR bin at ~0.6 mab as it varies with hydraulic parameters at ME30. (A) Shields parameter ( $\theta$ ) and (B) absolute velocity ( $U_1$ ) at 1 mab. Black dashed vertical lines: threshold for correlation coefficient R² computation. Threshold is ${\theta }_{\mathit{\beta cr}}$ in (A) and visually determined in (B) using the binned averages (black dots). R² and p-values are given on the bottom right of each panel. Blue line in (B): linear fitting.

4 Discussion

4.1 Sediment resuspension and transport in Cold Density Currents

The above analysis shows that, at 30-m depth, CDC velocities and backscattering echo levels are correlated when these velocities are ≳0.09 m s⁻¹. Previously, Fer et al. (2002b) reported elevated ADCP acoustic backscattering associated with CDCs (initiated by similar surface buoyancy flux $B$ as during event $E$ ) at 55-m depth, and a significant correlation between current velocities and backscattering intensity. They hypothesized that this correlation was due to sediment resuspension by the CDCs. This is confirmed by the present results at 30-m depth where, at ~1 mab, there is a strong correlation between high current velocities and the echo level (Figure 4B), which suggests resuspension. Echo level impulses, confined in the bottom dense layer by the strong stratification, are observed with the arrival of CDCs (Figure 3C). We also observed that resuspension bursts are more intense at the head of the CDCs, confirming laboratory (Maggi et al., 2024) and numerical (Kyrousi et al., 2018) observations. On the other hand, no indication of resuspension was found at mooring ME10 (at 10-m depth) on the edge of the shelf where CDC velocities were lower. Cold water steeply plunged at around this depth and CDCs were not in a dynamically balanced state, i.e., not yet fully developed, as at 30-m depth (Thorpe and Ozen, 2007) which explains the lower velocities.

Fer et al. (2002b) observed a strong acoustic backscattering signal above 5 mab, which indicates that transport of suspended sediment was well developed at 55-m depth (~850 m offshore) compared to the present measurements at 30-m depth (~450 m offshore) where resuspension impulses are short and vertically confined close to the bed. Furthermore, at 55-m depth, the height of maximum CDC velocity above the bed was between 5 and 8 mab, with an upper limit of the CDC between 16 and 25 m (Fer et al., 2002a; Thorpe and Ozen, 2007), indicating that the density current is much thicker at 55-m depth due to entrainment of ambient water. Therefore, the height of 5 mab where the higher backscattering was observed is still within the CDC’s internal dense layer where cross-shore velocities were of the same order of magnitude as those observed in the present study. By combining previous and present results obtained at this field site, it is concluded that, in CDCs: (1) sediment is not resuspended on the coastal shelf or on the shelf edge, even though the lakebed is steep there (30° at ME10), (2) resuspension is initiated further downslope, between 10 and 30-m depth, where the CDC is well-developed, and (3) resuspended sediments are present in a relatively thick layer at 55-m depth. These results indicate that at ~30-m depth, sediment transport in CDCs can exist. Furthermore, since the CDCs are strong, sediment transport can potentially reach the deep residual thermocline (Supplementary Figure 1) and subsequently cause efficient sediment focusing further offshore.

4.2 Parameterization of sediment resuspension

The analysis demonstrates that sediment can be intermittently resuspended and potentially transported within CDCs. The episodic nature of the observed resuspension indicates that high temporal and spatial resolution velocity measurements and robust characterization of velocity statistics in CDCs are required to adequately assess resuspension processes, as was found in similar lake studies (e.g., Churchill et al., 2004; Valipour et al., 2017). However, an analysis based on the Shields parameter concept, which is often applied in loose boundary layer studies, did not provide clear results.

The Shields parameter ( $\theta$ ) was found to be only weakly correlated with high acoustic echo levels (Figure 4A). However, when considering all CDCs during the winter season, velocity at 1 mab ( $U_1$ ) was found to be a good predictor (and better than $U_{\max }$ ) for the occurrence of higher echo levels and thus sediment resuspension (Figure 4B). The better performance of $U_1$ compared to $\theta$ is likely due to differences between the underlying assumptions of the Shields concept and the measured, variable CDC flow fields. It is difficult to estimate friction velocity correctly in flows with strong velocity variability on short time scales, as were observed in the CDCs. These variations can cause significant accelerations, which invalidate the assumption of uniform flow underlying the Shields parameter concept. In oscillating flows, Lorke et al. (2002) found that even with a perfect logarithmic profile fit, the estimation of the friction velocity might be biased because the Law-of-the-Wall region is sometimes confined to a thin (unresolved) near-bed region. A similar bias in CDCs, due to rapid velocity variability, could also have occurred in the present study. Furthermore, sediments may have been resuspended outside the measurement zone and been advected to ME30.

The analysis shows that resuspension from the sloping lateral bed in the littoral zone of Lake Geneva occurs frequently during winter cooling events that trigger CDCs. In the case of winter cascading (which is different from classical wind-driven resuspension events), resuspended particles are then potentially advected far downslope until the CDC waters reach the level of neutral buoyancy in the deep hypolimnion (Supplementary Figure 1). Due to their frequent occurrence (~25% of winter days), CDCs are expected to contribute to sediment focusing in Lake Geneva. Quantifying this contribution requires determining the suspended sediment concentration by monitoring turbidity and analyzing water samples.

5 Summary and conclusions

Sediment resuspension by Cold Density Currents (CDCs), generated by differential cooling, was investigated in the nearshore zone of Lake Geneva during winter 2017/2018 using high-resolution ADCP and temperature profile data. Acoustic backscattering recorded by ADCPs is an indicator of sediment resuspension.

This study revealed that:

• No resuspension occurred on the shallow coastal shelf (~6-m depth) where CDCs were initiated, and at 10-m depth, on the steep slope (30°) just beyond the shelf edge.

• CDCs flowed as impulses, and intermittent resuspension bursts were observed when CDC impulses arrived at the mooring located at 30-m depth, on a ~ 4.5° slope, ~450 m offshore, where CDCs were well developed and had acquired sufficient velocity.

• Resuspension of the fine local sediment (mean diameter $23\mathit{\mu m}$ ) was generally limited to a ~ 1.5-m thick near-bottom layer, consistent with the CDC stratification.

• Statistical analysis over the full-winter season indicates that CDCs occurred on 23 days (25% of the recording period), making them an important feature of winter circulation (and potentially of sediment focusing).

• Sediment is generally resuspended when CDC velocity at 1 m above the bed exceeds ~0.09 m s⁻¹, and sometimes at lower velocity in the head of CDC impulses.

• Individual downslope velocity profiles in the CDC bottom boundary layer were well described by a logarithmic profile, thus allowing determination of the Shields parameter via the bottom shear stress.

• CDC velocity measurements displayed marked temporal variability, which complicates the determination of a representative logarithmic mean profile needed to compute the Shields parameter.

• High ADCP echo levels did not always match Shields parameter values exceeding the threshold.

• By combining all the CDCs during the winter period, downslope velocity at ~1 m above the bed was found to be a better predictor for instantaneous sediment resuspension than the Shields parameter.

CDCs, which frequently occur in winter, are therefore an important process for transporting resuspended sediment from the lateral slope to deeper hypolimnetic waters. Since CDCs can travel far downslope when the lake is weakly stratified or completely destratified at the end of the winter, they can significantly impact nearshore and deep lake sediment dynamics. Moreover, CDCs can potentially redistribute polluted sediment downslope. Thus, the present findings are relevant for water resource management in littoral zones where contaminated outflows are often released into the shallow nearshore zone and drinking water is extracted from deeper depths in the same area. Similar nearshore dynamics can be expected to occur in other large, deep lakes under comparable conditions, further emphasizing the pertinence of the results of the present study.

Data availability statement

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found at: 10.5281/zenodo.7050849.

Author contributions

FM: Data curation, Conceptualization, Investigation, Writing – review & editing, Writing – original draft, Formal analysis. RR: Writing – review & editing, Investigation, Data curation. UL: Writing – review & editing, Conceptualization. TD: Writing – review & editing. KB: Writing – review & editing. DB: Writing – review & editing, Conceptualization. DAB: Conceptualization, Project administration, Writing – review & editing, Funding acquisition.

Funding

The author(s) declared that financial support was received for this work and/or its publication. This work was supported by The Bois Chamblard Foundation (Grant No. 532243). Open access funding by Swiss Federal Institute of Technology in Lausanne (EPFL).

Acknowledgments

Geneva long-wave irradiance data were provided by MeteoSwiss (the Swiss Federal Office of Meteorology and Climatology). We thank Htet Kyi Wynn and Benjamin Graf for assisting with the fieldwork and Violaine Piton for helping with the sediment analysis. We thank the two reviewers for their constructive comments and suggestions.

Conflict of interest

The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declared that Generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/frwa.2026.1693312/full#supplementary-material

References

Bloesch, J. (1995). Mechanisms, measurement and importance of sediment resuspension in lakes. Mar. Freshw. Res. 46, 295–304. doi: 10.1071/mf9950295

Crossref Full Text | Google Scholar

Boegman, L., and Stastna, M. (2019). Sediment resuspension and transport by internal solitary waves. Annu. Rev. Fluid Mech. 51, 129–154. doi: 10.1146/annurev-fluid-122316-045049

Crossref Full Text | Google Scholar

Churchill, J. H., Williams, A. J., and Ralph, E. A. (2004). Bottom stress generation and sediment transport over the shelf and slope off of Lake Superior’s Keweenaw peninsula. J. Geophys. Res. Oceans 109. doi: 10.1029/2003JC001997

Crossref Full Text | Google Scholar

CIPEL (2021). Rapports sur les études et recherches entreprises dans le bassin lémanique, Campagne 2020. Available online at: https://www.cipel.org/wp-content/uploads/catalogue/rapport-scientifique-2020-2021.pdf (accessed December 14, 2025)

Google Scholar

Cossu, R., Ridgway, M. S., Li, J. Z., Chowdhury, M. R., and Wells, M. G. (2017). Wash-zone dynamics of the thermocline in Lake Simcoe, Ontario. J. Great Lakes Res. 43, 689–699. doi: 10.1016/j.jglr.2017.05.002

Crossref Full Text | Google Scholar

de Durrieu Madron, X., Houpert, L., Puig, P., Sanchez-Vidal, A., Testor, P., Bosse, A., et al. (2013). Interaction of dense shelf water cascading and open-sea convection in the Northwestern Mediterranean during winter 2012. Geophys. Res. Lett. 40, 1379–1385. doi: 10.1002/grl.50331

Crossref Full Text | Google Scholar

Doda, T., Ramón, C. L., Ulloa, H. N., Wüest, A., and Bouffard, D. (2022). Seasonality of density currents induced by differential cooling. Hydrol. Earth Syst. Sci. 26, 331–353. doi: 10.5194/hess-26-331-2022

Crossref Full Text | Google Scholar

Doda, T., Ulloa, H. N., Ramón, C. L., Wüest, A., and Bouffard, D. (2023). Penetrative convection modifies the dynamics of downslope gravity currents. Geophys. Res. Lett. 50:e2022GL100633. doi: 10.1029/2022GL100633

Crossref Full Text | Google Scholar

Fer, I., Lemmin, U., and Thorpe, S. A. (2001). Cascading of water down the sloping sides of a deep lake in winter. Geophys. Res. Lett. 28, 2093–2096. doi: 10.1029/2000GL012599

Crossref Full Text | Google Scholar

Fer, I., Lemmin, U., and Thorpe, S. A. (2002a). Contribution of entrainment and vertical plumes to the winter cascading of cold shelf waters in a deep lake. Limnol. Oceanogr. 47, 576–580. doi: 10.4319/lo.2002.47.2.0576

Crossref Full Text | Google Scholar

Fer, I., Lemmin, U., and Thorpe, S. A. (2002b). Winter cascading of cold water in Lake Geneva. J. Geophys. Res. Oceans 107, 13–11. doi: 10.1029/2001JC000828

Crossref Full Text | Google Scholar

Géodonnées Etat de Vaud,. (2016a). Bathymétrie du Léman : MNT (grilles). Available online at: https://viageo.ch/catalogue/donnee/200617 (accessed December 14, 2025).

Google Scholar

Géodonnées Etat de Vaud,. (2016b). Bathymétrie du Léman : isobathes. Available online at: https://viageo.ch/catalogue/donnee/200618 (accessed December 14, 2025).

Google Scholar

Gould, H., and Budinger, T. (1958). Control of sedimentation and bottom configuration by convection currents, Lake Washington, Washington. J. Mar. Res. 17: 183–198. Available online at: https://elischolar.library.yale.edu/journal_of_marine_research/914 (accessed December 14, 2025)

Google Scholar

Graham, N. D., Bouffard, D., and Loizeau, J.-L. (2016). The influence of bottom boundary layer hydrodynamics on sediment focusing in a contaminated bay. Environ. Sci. Pollut. Res. 23, 25412–25426. doi: 10.1007/s11356-016-7715-9,

PubMed Abstract | Crossref Full Text | Google Scholar

Hawley, N. (2004). Response of the benthic nepheloid layer to near-inertial internal waves in southern Lake Michigan. J. Geophys. Res. Oceans 109, 1–14. doi: 10.1029/2003JC002128

Crossref Full Text | Google Scholar

Hilton, J., Lishman, J. P., and Allen, P. V. (1986). The dominant processes of sediment distribution and focusing in a small, eutrophic, monomictic lake. Limnol. Oceanogr. 31, 125–133. doi: 10.4319/lo.1986.31.1.0125

Crossref Full Text | Google Scholar

Horsch, G. M., and Stefan, H. G. (1988). Convective circulation in littoral water due to surface cooling. Limnol. Oceanogr. 33, 1068–1083. doi: 10.4319/lo.1988.33.5.1068

Crossref Full Text | Google Scholar

Kneller, B. C., Bennett, S. J., and McCaffrey, W. D. (1999). Velocity structure, turbulence and fluid stresses in experimental gravity currents. J. Geophys. Res. Oceans 104, 5381–5391. doi: 10.1029/1998JC900077

Crossref Full Text | Google Scholar

Kyrousi, F., Leonardi, A., Roman, F., Armenio, V., Zanello, F., Zordan, J., et al. (2018). Large eddy simulations of sediment entrainment induced by a lock-exchange gravity current. Adv. Water Resour. 114, 102–118. doi: 10.1016/j.advwatres.2018.02.002

Crossref Full Text | Google Scholar

Lohrmann, A. (2001). Monitoring sediment concentration with acoustic backscattering instruments. Available online at: https://www.nortekgroup.com/assets/documents/Monitoring-sediment-concentration-with-acoustic-backscattering-instruments.pdf (accessed December 14, 2025)

Google Scholar

Loizeau, J.-L., Dominik, J., Luzzi, T., and Vernet, J.-P. (1997). Sediment core correlation and mapping of sediment accumulation rates in Lake Geneva (Switzerland, France) using volume magnetic susceptibility. J. Great Lakes Res. 23, 391–402. doi: 10.1016/S0380-1330(97)70921-3

Crossref Full Text | Google Scholar

Loizeau, J.-L., Girardclos, S., and Dominik, J. (2012). Taux d’accumulation de sédiments récents et bilan de la matière particulaire dans le Léman (Suisse—France). Arch. Sci. 65:81. Available online at: https://www.unige.ch/sphn/Publications/ArchivesSciences/AdS%202004-2015/AdS%202012%20Vol%2065%20Fasc%201/081-092_08b_Loizeau_65.pdf (accessed December 14, 2025)

Google Scholar

Lorke, A., Umlauf, L., Jonas, T., and Wüest, A. (2002). Dynamics of turbulence in low-speed oscillating bottom-boundary layers of stratified basins. Environ. Fluid Mech. 2, 291–313. doi: 10.1023/A:1020450729821

Crossref Full Text | Google Scholar

Maggi, M. R., Negretti, M. E., Martin, A., Naaim-Bouvet, F., and Hopfinger, E. J. (2024). Experimental study of gravity currents moving over a sediment bed: suspension criterion and bed forms. Environ. Fluid Mech. 24, 1215–1233. doi: 10.1007/s10652-024-09998-6

Crossref Full Text | Google Scholar

Mikkelsen, O. A., and Pejrup, M. (2000). In situ particle size spectra and density of particle aggregates in a dredging plume. Mar. Geol. 170, 443–459. doi: 10.1016/S0025-3227(00)00105-5

Crossref Full Text | Google Scholar

Monismith, S. G., Imberger, J., and Morison, M. L. (1990). Convective motions in the sidearm of a small reservoir. Limnol. Oceanogr. 35, 1676–1702. doi: 10.4319/lo.1990.35.8.1676

Crossref Full Text | Google Scholar

Palandi de Mendosa, F., Schroeder, K., Miserocchi, S., Borghini, M., Giordano, P., Chiggiato, J., et al. (2023). Sediment resuspension and transport processes during dense water cascading events along the continental margin of the Southern Adriatic Sea (Mediterranean Sea). Mar. Geol. 459:107030. doi: 10.1016/j.margeo.2023.107030

Crossref Full Text | Google Scholar

Palanques, A., and Puig, P. (2018). Particle fluxes induced by benthic storms during the 2012 dense shelf water cascading and open sea convection period in the northwestern Mediterranean basin. Mar. Geol. 406, 119–131. doi: 10.1016/j.margeo.2018.09.010

Crossref Full Text | Google Scholar

Peeters, F., and Kipfer, R. (2009). “Currents in stratified water bodies 1: density-driven flows” in Encyclopedia of inland waters. eds. T. Mehner and K. Tockner. Second ed (Oxford: Elsevier), 525–533.

Google Scholar

Peng, N., Lemmin, U., Mettra, F., Reiss, R. S., and Barry, D. A. (2024). Deepwater renewal in a large, deep lake (Lake Geneva): identifying and quantifying winter cooling processes using heat budget decomposition. Water Resour. Res. 60:e2023WR034936. doi: 10.1029/2023WR034936

Crossref Full Text | Google Scholar

Reardon, K. E., Bombardelli, F. A., Moreno-Casas, P. A., Rueda, F. J., and Schladow, S. G. (2014). Wind-driven nearshore sediment resuspension in a deep lake during winter. Water Resour. Res. 50, 8826–8844. doi: 10.1002/2014WR015396

Crossref Full Text | Google Scholar

Reiss, R. S., Lemmin, U., Cimatoribus, A. A., and Barry, D. A. (2020). Wintertime coastal upwelling in Lake Geneva: an efficient transport process for Deepwater renewal in a large, deep lake. J. Geophys. Res. Oceans 125:e2020JC016095. doi: 10.1029/2020JC016095

Crossref Full Text | Google Scholar

Rowan, D. J., Kalff, J., and Rasmussen, J. B. (1992). Estimating the mud deposition boundary depth in lakes from wave theory. Can. J. Fish. Aquat. Sci. 49, 2490–2497. doi: 10.1139/f92-275

Crossref Full Text | Google Scholar

Shteinman, B., Eckert, W., Kaganowsky, S., and Zohary, T. (1997). “Seiche-induced resuspension in Lake Kinneret: a fluorescent tracer experiment,” in The interactions between sediments and water: proceedings of the 7th international symposium, Baveno, Italy 22–25 September 1996, eds. R. D. Evans, J. Wisniewski, and J. R. Wisniewski (Dordrecht: Springer Netherlands), 123–131.

Google Scholar

Soulsby, R. L., and Whitehouse, R. J. S. 1997. Threshold of sediment motion in coastal environments., in Pacific coasts and ports ‘97: Proceedings of the 13th Australasian coastal and ocean engineering conference and the 6th Australasian port and harbour conference; 1, 145–150

Google Scholar

Thorpe, S. A., and Ozen, B. (2007). Are cascading flows stable? J. Fluid Mech. 589, 411–432. doi: 10.1017/S0022112007007902

Crossref Full Text | Google Scholar

Turner, J. S. (1973). “Turbulent shear flows in a stratified fluid” in Buoyancy effects in fluids (Cambridge: Cambridge University Press), 127–164.

Google Scholar

Valipour, R., Boegman, L., Bouffard, D., and Rao, Y. R. (2017). Sediment resuspension mechanisms and their contributions to high-turbidity events in a large lake. Limnol. Oceanogr. 62, 1045–1065. doi: 10.1002/lno.10485

Crossref Full Text | Google Scholar

Valipour, R., Bouffard, D., and Boegman, L. (2015). Parameterization of bottom mixed layer and logarithmic layer heights in Central Lake Erie. J. Great Lakes Res. 41, 707–718. doi: 10.1016/j.jglr.2015.06.010

Crossref Full Text | Google Scholar