Coastal Dynamics 2013 SALINITY AND SUSPENDED SEDIMENT DYNAMICS IN RESPONSE TO FORCING CHANGES IN A SMALL MACROTIDAL ESTUARY (CHARENTE, FRANCE) Florence Toublanc1, Isabelle Brenon2, Thibault Coulombier3, Olivier Le Moine4 Abstract A three-dimensional numerical model was developed to reproduce the hydrodynamics and salinity dynamics of the macrotidal Charente estuary, located on the French Atlantic coast. The model uses horizontal structured grids and vertical sigma layers. Tidal elevations from a series of nested models and river flows are considered for the open boundary conditions. The roughness length and the turbulent scheme were adjusted in order to better fit the tidal elevation, currents, and salinity observed on different locations. The simulations give good results, with a decreasing accuracy when moving upstream. The validated model was used to simulate the impact of the river flow and the spring/neap tidal cycle on the salinity intrusion and the residual circulation. Sediment fluxes were also evaluated from currents simulated in realistic conditions of weather and river flow, associated with turbidity in-situ measurements. Key words: Estuary, saline intrusion, vertical stratification, sediment dynamics, numerical 3D modeling. 1. Introduction Estuarine tidal circulation, salinity and sediment transport processes are the result of complex interactions between coastal ocean and upland rivers. Water surface elevation and current velocities within the estuary result from non-linear interactions between the river flow and the incident open sea hydrodynamics at the mouth. The resultant hydrodynamic behavior of the estuary determines the salinity structure and the sediment transport. A classical two-layer circulation emerges from the combination of seaward transport induced by the river and a periodic tidal seaward/landward transport. In a macrotidal estuary, tide-induced turbulence can be large enough to mix water bodies and impede the formation of a stratified water column. Previous studies investigated numerically the incidence of forcing variations (tides, river flow and wind) on saline intrusion and vertical stratification. Prandle (2004) explained that for partially mixed estuaries, an increase in tidal currents provoked a decrease in density related mixing, but an increase in mixing due to tidal straining. Gong et al. (2011) showed that the salt intrusion length depends greatly on the river discharge. Sediment transport and suspended sediment concentrations are strongly dependent on the tides and the river discharge. Allen et al. (1980) demonstrated that the turbidity maximum in the Gironde estuary is moved upstream for low river discharge, and downstream during high discharge. This behavior was confirmed by several studies in the Gironde (Sottolichio et al., 2000), or in other macrotidal estuaries (Brenon et al., 1999; Uncles et al., 2006). Higher suspended sediment concentrations are expected for the highest tidal ranges, against low concentrations for neap tides. Strong vertical mixing due to fast tidal currents occurs during spring tides, eroding more sediment. During neap tides, tidal currents are weaker and the sediment is settling (Dyer, 1997). Combination of high river runoff and spring tides usually leads to high suspended sediment concentrations and seaward export (Castaing and Allen, 1981; Lopes et al., 1 UMR 7266 CNRS-LIENSs, University of La Rochelle, 2 rue Olympe de Gouges, La Rochelle 17000, France. [email protected] 2 UMR 7266 CNRS-LIENSs, University of La Rochelle, 2 rue Olympe de Gouges, La Rochelle 17000, France. [email protected] 3 UMR 7266 CNRS-LIENSs, University of La Rochelle, 2 rue Olympe de Gouges, La Rochelle 17000, France. [email protected] 4 IFREMER-LERPC, Center of La Tremblade, Ronce-les-bains, La Tremblade 17390, France. [email protected] 1707 Coastal Dynamics 2013 2006). Under these conditions, the turbidity maximum is located downstream and tidal induced mixing is enhanced, redistributing sediments in the water column that can be exported by high surface velocities (Castaing and Allen, 1981). Understanding the salinity and turbidity dynamics of the Charente estuary represent an important issue since the river is a source of fresh drinkable water for the area. Economic interests are also considered since local leisure activities and oyster farming are greatly dependent on the suspended sediment concentrations and the outflows directions (Ravail et al., 1988; Modéran et al.; 2010; Modéran et al., 2012). A field study, including a large set of in-situ data, was conducted, highlighting the strong effect of tides and river runoff on the turbidity maximum extension and position, and on the suspended sediment concentration (Coulombier et al., 2013). The aim of this study is to determine the impact of tides and river flow on the vertical stratification and horizontal distribution of salinity, on the residual circulation, and on the sediment fluxes at the river mouth, using a 3D-numerical model validated through in-situ data. 2. Material and methods 2.1. Study site The Charente estuary (45°96'N, 1°00'W) (Figure 1) is small, shallow, and under the influence of a semidiurnal tide. Tidal range can reach 6.5 meters at the river mouth, with a 5 meters mean level, which indicates a macrotidal regime. Mean river flow is estimated at 70 m3/s, reaching maximum values of 600 m3/s over the last decade. A dam is located at Saint-Savinien, 50 km from the mouth, and the river flows into the Marennes-Oléron bay, in the southern part of the Pertuis Charentais. The tidal dynamics of the area are characterized by an inversion of the tidal asymmetry over the spring/neap cycle: flood duration is shorter than ebb duration during spring tides and longer during neap tides (Toublanc et al., 2012). The sediment encountered is cohesive and the estuary is highly turbid, with a great influence on the sediment distribution of the Marennes-Oléron Bay (Ravail et al., 1988; Modéran et al., 2010; Modéran et al., 2012). Figure 1. Map of the Charente Estuary and its adjacent coastal area. The locations of the stations used to calibrate and validate the model are indicated. 1708 Coastal Dynamics 2013 2.2. 3D numerical model The 3D numerical model MARS3D is a finite differences model with sigma coordinates, solving the Navier-Stokes equations with the classic Boussinesq and hydrostatic hypotheses. It has been fully described by Lazure and Dumas (2008). The model is forced by atmospheric conditions (wind, air temperature, atmospheric pressure, cloud cover and relative humidity) over the entire domain, provided by the Météo-France model ARPEGE (0.5° and 6h resolution). Sea surface elevation is imposed at the open boundaries and is obtained from a series of five nested models of decreasing extensions and increasing spatial resolutions. The finest grid is represented horizontally by a 30 m uniform mesh of 1405 x 766 points, and vertically by 5 sigma levels. Daily discharges of the Boutonne and the Charente rivers are considered. High resolution bathymetric data were provided through several organizations (SHOM, Ifremer, EPTB Charente) with datasets from 2003, 2007 and 2010. A two equations k-kl turbulence closure model was chosen (Warner et al, 2005a). This model, close to the Mellor-Yamada level 2.5 scheme (1974, 1982), solves the vertical eddy viscosity and diffusivity by computing the turbulent kinetic energy and the turbulence length scale. The horizontal eddy viscosity and diffusivity are computed by taking the local shear conditions into account, using the Smagorinsky method (1963). Bed roughness height is set to 0.0001 m. 2.3. In-situ measurements Model validation is performed using in-situ data recorded for different periods and at different locations detailed on Figure 1. Continuous tide gauge data from Rochefort and the Aix island (from the REFMAR portal) and ADCP data (1200 kHz, 5 minutes bursts, February to April 2011) at the river mouth (Lupin) were provided for sea surface elevation and current speed validation. Simulated salinity levels were compared to datasets obtained from multi-parameter probes (YSI 6600V2) located at the mouth, surface (February to March 2011, 10 minutes sampling) and bottom (October to December 2012, 5 minutes sampling). In order to estimate sediment fluxes, turbidity data were also recorded at the mouth (approximately 2 meters from the bed) with YSI 6600V2 probe (6026 optic turbidimeter, 5 minutes sampling). Salinity and turbidity are laboratory calibrated using 12880Us.cm-1 and formazine 1000 and 4000 NTU standard solutions. 3. Results 3.1. Model validation Model validation was conducted over different periods of time, depending on the in-situ data available. Modeled water surface elevation, velocity currents and salinity concentrations were compared to observed data. For each variable, the mean absolute error (MAE) and the root mean square error (RMSE) were calculated. A skill parameter developed by Willmott (1981) and recently used in different estuarine dynamics modeling studies (Li and Boicourt, 2005; Ma et al., 2011; Warner et al., 2005b; Xing et al., 2012) was also considered to estimate the accuracy of the model: = 1 − ∑| − | ∑| − | + | − | (1) This parameter allows to take into account modeled and observed deviations around the observed mean to estimate the performance of the simulation. Perfect agreement between model and observations corresponds to a skill parameter of 1. All calculations are presented in Table 1. 1709 Coastal Dynamics 2013 Table 1. Model/data comparison statistics for water surface elevation, current velocity and salinity. Variable Location Water surface elevation Ile d'Aix 0.15 m 0.21 m 0.9937 Lupin 0.23 m 0.29 m 0.9904 Rochefort 0.37 m 0.45 m 0.9746 Lupin (averaged) 0.15;0.08 m/s 0.18;0.10 m/s 0.9762;0.9205 Rochefort (bottom) 0.11;0.14 m/s 0.14;0.18 m/s 0.9075;0.9283 Current velocity (u;v) Salinity MAE RMSE Skill (equation (1)) Lupin (surface) 3.24 psu 4.06 psu 0.9432 Lupin (bottom) 3.51 psu 4.64 psu 0.9261 3.1.1. Tidal elevation and currents Tidal elevation is compared to tide gauge data from two locations (Ile d’Aix, Rochefort), and ADCP data at the mouth (Lupin). Comparison for Lupin is presented in Figure 2. RMS errors and skill parameters (Table 1) suggest that the model reproduces well the water surface elevation, with a larger error for the most upstream location (Rochefort). However, double high waters are often not reproduced by the model. High waters are more accurate than low waters, both for the Lupin and Rochefort locations. At the mouth, this error can be related to the position of the ADCP, which was not fully immersed at low water during spring tides. Concerning the Rochefort station, Walther et al. (2007), showed that the presence of fluid mud could lead to a significant decrease of low water levels that is not reproduced by hydrodynamic models. This phenomenon would be consistent with the sediment dynamics of the Rochefort area where high suspended sediment concentrations lead to fluid mud deposition (Coulombier et al., 2013). Figure 2. Comparison of water surface elevation between model (solid) and observations (dots) at Lupin Comparison of ADCP and modeled depth-averaged currents at Lupin are presented in Figure 3. Some data are missing since the ADCP was not fully immersed at all times, particularly at low water during spring tides. Model results agree with observed data reasonably well, but the maximum currents are generally underestimated, especially for the meridional velocity. This observation is confirmed by the skill parameter (0.92), inferior to the one calculated for the zonal velocity (0.97). Bottom ADV and simulated velocities are compared at Rochefort. In this case, meridional velocities are better reproduced than zonal velocities (0.93 skill parameter against 0.91). Maximum current velocities are again globally underestimated, especially for the zonal velocity. For the two locations, simulated along-channel currents are then better simulated than lateral currents. Both for current velocities and for water surface elevation, simulated and observed data are very well in phase. 1710 Coastal Dynamics 2013 Figure 3. Comparison of zonal (U) and meridional (V) depth-averaged current velocities between model (solid) and observations (dots) at the river mouth. 3.1.2. Salinity Surface salinity concentrations at the mouth are presented in Figure 4. Both the RMS error (4 psu) and the skill parameter (0.94) show that the model reproduces well the surface salinity (Table 1). The peaks due to low or high water are almost perfectly in phase for the whole simulation. Transitions between spring and neap tides are the dominant source of error. Between the 26th of February and the 2nd of March, surface salinity is particularly overestimated by the model. This could be explained by the precipitations that occurred in the area at this time, combined with a slight peak in wind velocities (around 30 km/h). These atmospheric conditions were not observed for the second neap tide (around the 13th of March). Bottom salinity at the mouth is also well reproduced with a skill parameter of 0.92 (Table 1). Figure 4. Comparison of surface salinity concentrations between model (solid) and observations (dots) at the river mouth 3.2. Response to tidal and river forcing changes 3.2.1. Saline intrusion and stratification After calibration and validation, the model is used to estimate the impact of tides and river flow on the 1711 Coastal Dynamics 2013 Depth (m) salinity dynamics of the estuary. Two river flows were considered, over a spring/neap tidal cycle: 5 m3/s and 400 m3/s. For both simulations, realistic meteorological conditions were considered, respectively representative of summer and winter periods. Transect salinity distributions along the estuary are presented in Figure 5 for low and high waters of the strongest spring tide and the weakest neap tide encountered during the simulated period. LW-NT LW-NT HW-NT HW-NT LW-ST LW-ST HW-ST HW-ST Distance from the mouth (km) Figure 5. Salinity transects along the estuary, 5m3/s (left) and 400 m3/s (right). (LW-NT): low water, neap tide. (HW-NT): high water, neap tide. (LW-ST): low water, spring tide. (HW-ST): high water, spring tide. Vertical and horizontal salinity distributions are evaluated in Table 2, in function of the tidal and river forcing. Stratification at Lupin and Rochefort are presented as the difference between bottom and surface salinity. The salinity intrusion limit is given by the distance from the mouth at which the bottom salinity is becoming inferior to 1 psu. Calculations are made for the same conditions as for Figure 5. Table 2. Salinity stratification and salinity intrusion in function of tidal and river forcing. River flow (m3/s) Bottom to surface salinity difference (psu) : Lupin Bottom to surface salinity difference (psu) : Rochefort LW 34.36 ≈0 7.9 HW 34.72 ≈0 12.1 LW ≈0 ≈0 0.4 HW 25.41 ≈0 12.1 LW 0.66 1.57 33.8 HW 0.32 4.21 41.1 LW 0.05 1.60 40.2 HW 0.01 2.13 43.8 Tidal conditions Neap Tide 400 Spring Tide Neap Tide 5 Spring Tide Salinity intrusion limit (km) As expected, saline intrusion length is maximum for the highest tidal range and the lowest river flow (Figure 5, Table 2). For a 5m3/s river flow, the movement of this intrusion between low and high water is higher for neap tide (7.3 km) than for spring tide (3.6 km). The intrusion variation between neap and spring tides is about 10 km. Salinity stratification at the mouth is inferior to 1 psu both for spring and neap tide. Transects also show the channel weak stratification for low river input (Figure 5). Under high river flow conditions, salinity intrusion is moved downstream (12.1 km from the mouth for 1712 Coastal Dynamics 2013 neap and spring tides). The intrusion limit is almost pushed outside of the estuary at low water for spring tide. Its extension is higher for spring tide (11.7 km), but smaller for neap tide (4.2 km). Salinity stratification is very strong at the mouth, except during spring tide at low water. This is simply because there is only fresh water in the water column in the channel, which is also the case at Rochefort for all tidal conditions. Simulations show that salinity can locally be higher (up to 5 psu) on the sides of the channel at the mouth, probably because of the lower dynamics of the tidal flats. Strong vertical stratification for high runoff is confirmed by transects presented in Figure 5. For both runoff conditions, neap tide stratification is equivalent or higher than spring tide stratification. 3.2.2. Residual circulation at the mouth Tidally averaged current velocities are calculated at the river mouth to determine the tidal and fluvial influence on the residual circulation. Under low river flow conditions, residual velocities are small and decrease almost linearly from surface to bottom. They do not exceed 8 cm/s for neap tide and 19 cm/s for spring tide. Values decrease very rapidly in the first part of the water column: 72% and 75% drops, respectively for neap and spring tide, between the first two layers. With a 400 m3/s river flow, surface residual velocities reach 53 cm/s during spring tide and 48 cm/s during neap tide. Vertical distribution is more linear, especially for spring tides. For all cases, residual circulation is oriented seaward at the surface. At the bottom, residual velocities are all inferior to 8 cm/s. 3.3. Sediment dynamics Turbidity data at the Lupin station are available from October to December 2012. In order to estimate sediment fluxes at this station, the current velocities were simulated, with real atmospheric and runoff conditions. Two floods are observed over this period of time, with respectively 100 m3/s and 350 m3/s runoffs. Turbidity data are converted, through calibration, to obtain suspended sediment concentrations in g/L. Current velocities are extracted from the model to obtain suspended sediment fluxes for each time step (flood 1: Figure 6a, flood 2: Figure 6b). Negative values indicate sediment exporting, positive values indicate sediment importing. Net sediment transport is calculated over a tidal cycle. Since the turbidity variability (vertically and horizontally) observed on the field is significant, the choice was made to not integrate these values on the water column or on the channel cross-section. It is also important to notice that calculated fluxes remain estimations since the turbidity sensors can only measure up to 4000 NTU. As described by Coulombier et al. (2013), this limit is exceeded very often, more particularly during high runoff and spring tides. (a) (b) Figure 6. Suspended sediment fluxes (kg/m²/s) and net transport per tide (ton/m²/tide) at the mouth in function of the river flow and tidal range. (a): flood 1 (≈100 m3/s). (b): flood 2 (≈350 m3/s). 1713 Coastal Dynamics 2013 Spring tide conditions provoke higher sediment transport, due to higher current velocities. For both cases, higher net sediment transport occurs during spring tide, respectively 12.9 and -16.1 tons/m²/tide. For flood 1 (Figure 6a), four phases can be observed. First, with spring tides and a very low river discharge (≈20 m3/s), net sediment transport is negative, indicating seaward transport (15th-18th of October). As the river flow increases (20 to 100 m3/s) and the tidal range decreases, sediment is imported (19th-23rd of October). Net transport is approaching zero under neap tide conditions, with the river discharge reaching its maximum of 105 m3/s on October 24th and dropping to 76 m3/s on October 26th. Finally, tidal range increases and river flow decreases, leading to landward sediment transport. For flood 2 (Figure 6b), such phases are more difficult to observe since net sediment transport values are less linearly distributed. Between the 13th and the 17th of December, sediment tends to be exported, under spring tide conditions and average river flow (≈60 m3/s). Tidal range decreases, and river discharge increases (60 to 210 m3/s) between the 17th and the 20th, leading to net sediment import. Sediment transport is then directed seaward, as the river flow keeps on rising and the tides approach neap conditions (20th-23rd of December). At the runoff peak (December 23rd to 26th) and during neap tide, sediments seem to be imported again. In both cases, net sediment transport is sensitive to river flow variations, but is strongly modulated by the spring-neap tidal cycle. According to field observations, salinity concentrations at the same location remained driven by the tide, meaning that the river flow was not large enough, or not for a sufficient period of time, to eject the turbidity maximum over several tidal cycles. However, salinity variations between high and low tide are higher during flood 2 (26 psu against 18 psu for flood 1), suggesting that the turbidity maximum is moved downstream, and could be oscillating at the mouth, in function of the tides. 4. Discussion 4.1 Model validation A three-dimensional finite differences model was developed to reproduce the hydrodynamics of the Charente estuary, and to assess the impact of tidal and runoff variations on estuarine dynamics. The model performance was first evaluated by calculating the mean absolute error (MAE), the root mean square error (RMSE), and a skill parameter developed by Willmott (1981) describing the correspondence between the observed and predicted deviations around the observed mean. This series of statistics (Table 1) shows that the model gives a good representation of the water surface elevation, the current velocities and the salinity concentration in the Charente estuary (skill > 0.9 for all parameters). However, modeled low water levels at Rochefort are not very well reproduced (Table 1). This error could be linked to the presence of deposited fluid mud observed in the area (Coulombier et al., 2013), inducing very low bottom friction. Taking into account the variation of friction due to sediment dynamics seems then essential to improve the model accuracy, as shown by Walther et al. (2007). Preliminary results obtained from a test with very low friction showed an improvement of errors on the water level at Rochefort (from 0.45 m RMSE to 0.34 m). Bathymetric uncertainties also need to be considered as a source of error. The salinity response to atmospheric forcing (Figure 4) could be enhanced with higher resolution data. Li et al. (2005) also showed that turbulence mixing schemes often fails to reproduce correctly salinity stratification under high runoff conditions. To improve model predictions, they suggest that the background diffusivity should not be a constant but a time and stratification dependent variable. 4.2 Estuarine circulation and mixing dependence on fluvial and tidal forcing The effect of the spring-neap tidal cycle and the river runoff on salt intrusion and stratification is investigated in the Charente estuary (Table 2). As shown previously by several studies (Liu et al., 2007, Azevedo et al., 2010, Gong and Shen, 2011), river flow influences greatly the saline intrusion length in the estuary. For a very low runoff, the intrusion limit (defined as the 1 psu limit) is localized between 34 and 44 km from the river mouth, highlighting the low dynamics of this configuration, already observed in other studies (Bowen and Geyer, 2003; Gong and Shen, 2011). Under high runoff conditions, salt is exported 1714 Coastal Dynamics 2013 towards the Marennes-Oléron Bay and stratification at the mouth can be very high (up to 34 psu). Residual circulation calculations confirm this behavior since surface seaward velocities can be six times higher than with low runoff. The surface layer is dominated by the river flow. Salinity at the Aix island can decrease below 10 psu, confirming the strong influence of the river on the surrounding area and its ecosystem. Spring tide salinity stratification is equivalent or weaker than neap tide stratification, especially with high runoff. Calculated residual velocities are vertically well distributed under these conditions, showing that tidal mixing induced by higher velocities is more important during spring tide. During neap tides, tidal straining decreases and velocities are reduced, preventing the turbulent mixing from developing (Lewis, 1997; Thain et al., 2004). The movement of the salinity intrusion limit between low and high water is higher for spring tide and high runoff, but also higher for neap tide and low runoff. For the weaker fluvial forcing, this limit is located far upstream (around 40 km), where the tidal influence is lower than when the river flow is higher and the intrusion limit is located close to the mouth. However, it is necessary to notice that tidal range difference between the spring and neap tides considered is higher for the 400 m3/s case than for the 5 m3/s case. A stronger spring tide would probably lead to a deeper upstream penetration of salinity and a larger difference between low and high water. 4.3 Response of sediment transport to tides and river flow As expected, sediment transport is enhanced by the highest tidal ranges, which provoke higher velocities. Sediment fluxes are then more important for two reasons: higher velocities means more transport per unit of time, but also more straining on the river bed, leading to erosion and higher suspended sediment concentrations. Turbulence and mixing are also more important, whereas neap tide weaker currents allow sediments to settle (Dyer, 1997). Sediment fluxes semi diurnal and fortnightly variability observed in Figure 6 demonstrates that sediment dynamics are strongly determined by tidal currents. Net sediment export from the estuary requires a river flow large enough, and over a significant period of time. According to field observations and results shown in Figure 6, these conditions were not satisfied during the two periods simulated, since the turbidity maximum does not seem to be ejected from the estuary. Both runoff peaks also occurred during neap tides, when suspended sediment concentrations are lower. As shown by Geyer et al. (2001), net sediment export can be more determined by the timing between high runoff events and the spring-neap tidal cycle than by the intensity of the river flow. However, salinity variations at the mouth, over a tidal cycle, are higher for the strongest river flow (26 psu for 350 m3/s against 18 psu for 100 m3/s), suggesting that the turbidity maximum, which is generally linked to the salinity intrusion (Uncles and Stephens, 1993, Lopes et al., 2006), is moved downstream under high runoff conditions. For both cases, net sediment export is observed for spring tides and low (20 m3/s) or average (60 m3/s) river flow. During spring tides, the flood is shorter than the ebb but flood current velocities do not exceed ebb velocities by more than 10 cm/s in most cases. Ebb currents can even be stronger. Since the ebb is longer, this could lead to higher sediment export. Such export under spring tide and low runoff conditions was also observed by Kitheka et al. (2005). However, since turbidity probes saturated both during ebb and flood, suspended sediment concentrations could be underestimated, more particularly during floods, leading to errors in sediment transport estimations. Sediment import observed from the 23rd of December (Figure 6b) could be explained by the strengthening of flood tidal currents observed at this time. Since the turbidity probe is located near the bottom, sediment erosion from the bed due to stronger currents, combined with higher flood velocities, could explain this behavior. Near-bed dynamics at the mouth are mostly determined by the tides, explaining why this phenomenon would occur, even if the river runoff is high. Coulombier et al. (2013), showed that more than half of temporal variations of sediment transport at the mouth is explained by the tidal regime. Further investigation is needed to determine when the turbidity maximum is exported in the MarennesOléron Bay, possibly affecting oyster farming. Incorporation of sediment transport in the model is essential, because of the vertical and horizontal variability of velocity currents and suspended sediment concentrations, but also because of the 4000 NTU limit of probes. 1715 Coastal Dynamics 2013 5. Conclusion The 3D numerical model developed for this study gives a good representation of the Charente estuary hydrodynamics, especially at the river mouth. Different fluvial and tidal conditions can be tested to determine their effect on the salinity dynamics and the residual circulation. Results show that the estuary is more stratified during neap tides, due to weaker velocities. Particularly high runoff enhances stratification, the surface layer of the estuary being dominated by the river. Saline intrusion is maximum for the highest tidal range, when the river input is small. Sediment transport is strongly related to the semi-diurnal and spring-neap tidal cycle. Net sediment export can occur for high river flow, more especially if this event is consistent with spring tides. Sediment transport will be included in the model for future research, allowing further investigations on the turbidity maximum formation and movement in function of various forcing. Acknowledgements The authors gratefully acknowledge funding from the Conseil Général of Charente Maritime, the CNRS, the FEDER and the University of La Rochelle. The REFMAR portal is also acknowledged for the data provided, as well as Ifremer for the bathymetry and the MARS3D code. This study would not have been possible without the technical and data support from the Ifremer-LERPC Laboratory. The authors thank particularly Nicolas Lachaussée, Philippe Pineau, Christophe Arnaud, Jean-Michel Chabirand, Florence Cornette, Philippe Geairon, Stéphane Robert and Jean-Luc Seugnet for their precious help on the field. References Allen, G.P., Salomon, J.C., Bassoullet, P., Du Penhoat, Y., De Grandpre, C., 1980. Effects of tides on mixing and suspended sediment transport in macrotidal estuaries, Sedimentary Geology, 26: 621-640. Azevedo, I. C., Bordalo, A. a, and Duarte, P. M., 2010. Influence of river discharge patterns on the hydrodynamics and potential contaminant dispersion in the Douro estuary (Portugal). Water research, 44(10): 3133–46. Brenon, I., Le Hir, P., 1999. Modeling the turbidity maximum in the Seine Estuary (France): identification of formation processes, Estuarine Coastal and Shelf Science, 49: 525-544. Bowen, M. M. and Geyer, R., 2003. Salt transport and the time-dependent salt balance of a partially stratified estuary. Journal of Geophysical Research, 108(C5): p.3158 Castaing, P. and Allen, G.P., 1981. Mechanisms controlling seaward escape of suspended sediment from the Gironde: a macrotidal estuary in France, Marine Geology, 40(1-2): 101-118. Coulombier, T., Toublanc, F., Brenon I., 2013. Seasonal monitoring of sediments dynamics in a highly turbid estuary (Charente estuary, France): source and sink of the turbidity maximum, Proceedings of the 7th International Conference on Coastal Dynamics, Arcachon. Dyer, K.R, 1997. Estuaries: a physical introduction, 2nd edition, John Wiley & Sons, 195 pp. Geyer, W. R., Woodruff, J. D., and Traykovski, P., 2001. Sediment transport and trapping in the Hudson River Estuary. Estuaries, 24(5): 670–679. Gong, W., and Shen, J., 2011. The response of salt intrusion to changes in river discharge and tidal mixing during the dry season in the Modaomen Estuary, China, Continental Shelf Research, 31: 769-788. Kitheka, J. U., Obiero, M., and Nthenge, P., 2005. River discharge, sediment transport and exchange in the Tana Estuary, Kenya. Estuarine, Coastal and Shelf Science, 63(3): 455–468. Lazure, P. and Dumas, F., 2008. An external–internal mode coupling for a 3D hydrodynamical model for applications at regional scale (MARS). Advances in Water Resources, 31(2): 233-250. Lewis, R., 1997. Dispersion in Estuaries and Coastal Waters. John Wiley, Chichester, 312 pp. Li, M., Zhong, L. and Boicourt, W. C., 2005. Simulations of Chesapeake Bay estuary: Sensitivity to turbulence mixing parameterizations and comparison with observations. Journal of Geophysical Research, 110: C12004. Liu, W.-C., Chen, W.-B., Cheng, R. T., Hsu, M.-H., and Kuo, A. Y., 2007. Modeling the influence of river discharge on salt intrusion and residual circulation in Danshuei River estuary, Taiwan. Continental Shelf Research, 27(7): 900– 921. Lopes, J.F., Dias, J.M., Dekeyser, I., 2006. Numerical modeling of cohesive sediments transport in the Ria de Aveiro Lagoon, Portugal, Journal of Hydrology, 319: 176-198. Ma, G., Shi, F., Liu, S. and Qi, D., 2011. Hydrodynamic modeling of Changjiang Estuary: Model skill assessment and large-scale structure impacts. Applied Ocean Research, 33(1): 69-78. 1716 Coastal Dynamics 2013 Mellor, G. L., and Yamada, T., 1974. A hierarchy of turbulence closure models for planetary boundary layers. Journal of the Atmospheric Sciences, 31: 1791–1806. Mellor, G. L., and Yamada, T., 1982. Development of a turbulence closure model for geophysical fluid problems. Reviews of geophysics and space physics, 20(4): 851–875. Modéran, J., Bouvais, P., David, V., Le Noc, S., Simon-Bouhet, B., Niquil, N., Miramand, P. and Fichet, D., 2010. Zooplankton community structure in a highly turbid environment (Charente estuary, France): Spatio-temporal patterns and environmental control. Estuarine, Coastal and Shelf Science, 88: 219–232. Modéran, J., David, V., Bouvais, P., Richard, P. and Fichet, D., 2012. Organic matter exploitation in a highly turbid environment: Planktonic food web in the Charente estuary, France. Estuarine, Coastal and Shelf Science, 98: 126– 137. Prandle, D., 2004. Saline intrusion in partially mixed estuaries, Estuarine Coastal and Shelf Science, 59: 385-397. Ravail, B., Heral, M., Maestrini, S., and Robert, J.-M. 1988. Incidence du débit de la Charente sur la capacité biotique du bassin ostréicole de Marennes-Oléron, Journal de Recherche Océanographique, 13: 48–52. Smagorinsky, J. (1963). General circulation experiments with the primitive equations I: The basic experiment. Monthly Weather Review, 91(3): 99–164. Sottolichio, A., Le Hir, P., Castaing, P., 2000. Modeling mechanisms for the turbidity maximum stability in the Gironde Estuary, Coastal and Estuarine Fine Sediment Processes, 373-386. Thain, R. H., Priestley, A. D., & Davidson, M. A., 2004. The formation of a tidal intrusion front at the mouth of a macrotidal, partially mixed estuary: a field study of the Dart estuary, UK. Estuarine, Coastal and Shelf Science, 61(1): 161–172. Toublanc, F., Brenon, I., Le Moine, O., 2012. Inversion de l’asymétrie de la marée au cours d’un cycle morte-eau viveeau, dans l’estuaire de la Charente (France). Actes des Journées Nationales Génie Côtier - Génie Civil 2012, 141148. Uncles, R. J., and Stephens, J. A., 1993. The Freshwater-Saltwater Interface and Its Relationship to the Turbidity Maximum in the Tamar Estuary, United Kingdom. Estuaries, 16(1): 126–141. Uncles, R.J., Stephens, J.A., Harris, C., 2006. Runoff and tidal influences on the estuarine turbidity maximum of a highly turbid system : The upper Humber and Ouse Estuary, UK, Marine Geology, 235: 213-228. Walther, R., Bertrand, O., Rieu, J., Hamm, L., 2007. Modélisation tridimensionnelle de la salinité et de la turbidité dans l’estuaire de la Loire: couplage des processus, La Houille Blanche, 4:47 :55. Warner, J. C., Sherwood, C. R., Arango, H. G., and Signell, R. P., 2005a. Performance of four turbulence closure models implemented using a generic length scale method. Ocean Modelling, 8: 81–113. Warner, J. C., Geyer, W. R. and Lerczak, J. A., 2005b. Numerical modeling of an estuary: A comprehensive skill assessment. Journal of Geophysical Research, 110: C05001. Willmott, C. J., 1981. On the validation of models. Physical Oceanography, 2: 184-194. Xing, Y., Ai, C., and Jin, S., 2012. A three-dimensional hydrodynamic and salinity transport model of estuarine circulation with an application to a macrotidal estuary. Applied Ocean Research, 39: 53-71. 1717 Coastal Dynamics 2013 1718
© Copyright 2026 Paperzz