1. Introduction
Estuaries are dynamically complex systems that are important from both an economical and ecological point of view. To be able to address issues related to these aspects, a good understanding of the physical processes resulting in the estuarine exchange flow is fundamental, as this flow drives long-term transport of sediment and other suspended particle matter in estuaries. Since the exchange flow is significantly affected by the salt dynamics, it is key to improve our understanding of the essential three-dimensional processes that result in the spatially and temporally varying salinity fields.
Salt transport processes can vary significantly from estuary to estuary, since estuaries are subject to different forcing conditions and have greatly varying geometries and bathymetries (Fischer 1972). The classical subtidal salt balance, in which the river-induced seaward transport is balanced by the landward transport due to the longitudinal–vertical gravitational circulation and dispersion, is extensively described in literature (see, e.g., Hansen and Rattray 1965; MacCready 2004; MacCready and Geyer 2010). In this way, both lateral processes and processes that vary on the tidal time scale are parameterized. This classical balance is challenged in partially to well-mixed estuaries where gravitational circulation is relatively weak (Jay and Smith 1990) and other processes are likely to dominate. In well-mixed estuaries, for instance, the tidal oscillatory transport due to temporal correlations on the tidal time scale between the longitudinal velocity and salinity can be the predominant salt import mechanism (McCarthy 1993; Wei et al. 2016). The tidal oscillatory salt transport can also be significant in partially stratified estuaries such as the Hudson estuary during neap tides (Chen et al. 2012; Wang et al. 2015). Even in salt wedge estuaries such as the Merrimack River, the landward salt transport is mainly associated with the tidal oscillatory processes rather than with the steady sheared gravitational circulation (Ralston et al. 2010). This implies the potential importance of considering residual salt transport mechanism due to tidal processes (Hughes and Rattray 1980; Geyer and MacCready 2014).
Apart from the importance of processes on the tidal time scale, the transverse circulation (also known as secondary estuarine circulation) can be of great importance for the landward salt transport (Fischer 1972; Dyer 1974; Smith 1977). Significant transverse circulations are observed in well-mixed, partially mixed, and stratified estuaries (Nunes and Simpson 1985; Becherer et al. 2015; Valle-Levinson et al. 2000; Lacy and Monismith 2001). These circulations are usually associated with lateral variations in bathymetry (Nunes and Simpson 1985; Valle-Levinson et al. 2000), Coriolis force (Lerczak and Geyer 2004; Valle-Levinson 2010), lateral density gradients (Huijts et al. 2006, 2011), and curvature-induced centrifugal forces (Lacy and Monismith 2001; Buijsman and Ridderinkhof 2008). Transverse circulations are also affected by the estuarine width-to-depth ratio (Schulz et al. 2015) and temporal/lateral variations of mixing (tidal and lateral straining; see Burchard et al. 2011; Becherer et al. 2015).
To investigate the influence of transverse processes on estuarine salt transport, the salt flux is usually decomposed into different contributions associated with transverse/vertical steady or oscillatory shear, using velocities and salinities from measurements (Fischer 1972; Dyer 1974; Hughes and Rattray 1980; West et al. 1990; Guymer and West 1992) or numerical models (Oey et al. 1985; Ralston and Stacey 2005; Ralston et al. 2010). This method, however, strongly depends on the quality/quantity of the measurements and the resolution and accuracy of numerical models. Moreover, it is difficult to relate the shear-induced salt transport to a specific physical process since the shearing itself can be a result of the interplay of many different processes.
To unravel the dominant salt transport mechanisms, idealized models are appropriate tools, as they are specifically developed to investigate the influence of each physical mechanism in isolation. This type of model has been successfully used to gain insights into the influence of lateral processes on water motion (Nunes and Simpson 1985; Li and O’Donnell 1997; Wong 1994; Valle-Levinson et al. 2003; Kumar et al. 2016) and lateral sediment trapping (Huijts et al. 2006, 2009, 2011). Existent idealized models aiming at understanding the salt dynamics (explicitly including processes at the tidal time scale), however, are limited to width-averaged models (McCarthy 1993; Wei et al. 2016). In these models, the coupled width-averaged water motion and salinity are dynamically resolved while the lateral processes are parameterized as diffusion. Since these models do not resolve the lateral dynamics, the relative importance of the transverse circulation on the residual salt balance are not investigated.
In this paper, we therefore develop a three-dimensional idealized model by extending the model of Wei et al. (2016). This model is specifically aiming at dynamically resolving the temporal and spatial (both longitudinal and lateral) variability of the water motion and salinity (i.e., prognostic in salinity, including lateral processes). With this model, the influence of each physical process on the residual salt balance can be systematically investigated. As a first step, the estuaries are assumed to be well mixed, that is, the top-to-bottom salinity difference is one order of magnitude smaller than the bottom salinity, and the effects of tidal straining are assumed to be small. The nonlinear system of the three-dimensional water motion and salinity is solved at the tidal time scale, employing a perturbation method to analytically obtain the vertical structures of the physical quantities and a finite-element method to calculate their horizontal variations. An iterative approach is used to calculate the coupled gravitational circulation and salinity field. Using this model, the influence of the longitudinal/transverse processes on the residual salt transport can be investigated, and their sensitivity to parameters/forcing conditions can be systematically studied.
The paper is structured as follows: The equations governing the coupled water motion and salinity (see section 2a) and the solution method (see section 2b) are introduced in section 2. In section 3, the salt dynamics for the default experiment (representative for the Delaware estuary; see section 3a) and the influences of estuarine convergence, bathymetry, and Coriolis force on salt transport mechanisms (see section 3b) are investigated. In section 4, the salt transport induced by the tidal advective diffusion (see section 4a) and gravitational circulation (see section 4b) are analyzed considering the influence of the estuarine convergence, Coriolis, and bathymetry, respectively. Conclusions are drawn in section 5.
2. Model description
a. Governing equations
To investigate the salt dynamics and transport mechanisms in tidal estuaries, an idealized three-dimensional model is developed for estuaries that are assumed to be tidally dominated and well mixed. The shape of the estuary (i.e., its geometry) is arbitrary (see Fig. 1). The seaward boundary ∂SΩ is connected with the open sea and forced by tides. A weir is assumed to be located at the landward boundary denoted by ∂RΩ, where a river flows into the estuary. The closed boundaries ∂CΩ are impermeable. The undisturbed water level is located at z = 0, and the free-surface elevation is denoted by z = η. The estuarine bed is located at z = −H(x, y), with H an arbitrary function of the horizontal coordinates (x, y).
A three-dimensional sketch of an estuary, with x and y the horizontal coordinates, and z the vertical coordinate, positive in the upward direction. The seaward, river, and closed boundaries are denoted by ∂SΩ, ∂RΩ, and ∂CΩ, respectively. The free-surface elevation and the estuarine bottom are located at z = η and z = −H(x, y), respectively. A sketch for an estuary with a more general geometry and bathymetry can be found in Kumar et al. (2016) in their Fig. 1.
Citation: Journal of Physical Oceanography 47, 7; 10.1175/JPO-D-16-0247.1
b. Solution method
1) Perturbation method
2) Numerical method
3. Results
In section 3a, the salt dynamics for the default experiment, representative for the Delaware estuary, will be discussed. In section 3b, the influence of estuarine convergence, Coriolis effects, and estuarine bathymetry on salt dynamics are investigated using three dedicated experiments.
a. Salt dynamics for the default experiment
Default parameters, representative of the Delaware estuary.
1) Tidally averaged salinity field and longitudinal salt transport
The tidally averaged salinity is prescribed to be 31 psu at the seaward side, and it decreases to 2 psu at ~120 km inside the estuary, as shown by the background colors in Fig. 2, where the white color indicates a salinity of ~31 psu and black indicates a vanishing salinity. Hence, the salt intrusion length Ls, which in this paper is defined as the distance between the estuarine mouth to the location where the cross-sectionally averaged salinity is 2 psu, is 120 km. This salt intrusion length agrees well with the observed salinity profile under low river discharge (Kuijper and Van Rijn 2011; Wei et al. 2016), confirming that the longitudinal dependency of the horizontal diffusivity Kh, as given in Eq. (51), is appropriate for the Delaware estuary. A lateral variation of S0 (of ~1 psu) is found, with larger tidally averaged salinities on the left shoals than those on the right, looking seaward. This lateral distribution qualitatively agrees with the observation of Wong (1994) near the mouth of the Delaware Bay. Furthermore, the top-to-bottom salinity difference (below 0.5 psu) is much smaller than the bottom salinity during the tidal cycle (not shown), justifying the well-mixed assumption.
(a) Tidally averaged salinity distribution S0(x, y) (see gray colors and dotted lines) and streamlines of the depth-integrated salt transport. Solid lines indicate counterclockwise residual salt transport, and dashed lines indicate clockwise salt transport (looking into the paper). (b) The cross-sectionally integrated along-channel salt transport due to tidal advective diffusion (TASF), gravitational circulation (GC), river discharge (RD), tidal rectification (AD), stress-free surface (NS), tidal return flow (TRF), and prescribed diffusion (DIF). Positive values denote landward salt transport, and negative values denote seaward salt transport. The total residual salt transport over the cross section is zero (dotted line).
Citation: Journal of Physical Oceanography 47, 7; 10.1175/JPO-D-16-0247.1
To visualize the residual salt transport pattern, streamlines of the depth-integrated salt transport are used. Since the depth-integrated residual salt transport is divergence free [see Eq. (27)], the amount of salt transported through any area connecting the same streamlines is constant. The streamlines of the residual salt transport are shown in Fig. 2a, where solid lines indicate counterclockwise salt transport, and dashed–dotted lines indicate clockwise salt transport (looking into the paper). It is found that, in the idealized Delaware estuary, the salt is mostly transported counterclockwise, meaning the salt is transported into the estuary from the left side of the estuary and out of the estuary from the right side of the deeper channel looking seaward.
Since the salt transport between the same streamlines is constant, large distances between these streamlines indicate small (normal) salt transports, and small distances indicate large transports. As shown in Fig. 2a, the distance between the two neighboring streamlines first increases from the left bank to the midaxis (at y = 0) and then decreases sharply toward the right bank. This suggests an interesting steady-state salt transport pattern: the salt is slowly transported landwards from the left side of the mouth toward the midaxis and gets quickly transported out of the estuary on the right of the deeper channel. To evaluate the relative importance of various longitudinal salt transport mechanisms, the along-channel salt transport at each cross section is calculated using Eq. (48). As shown in Fig. 2b, the seaward salt transport due to river discharge (line with upside-down triangles) is mainly balanced by the landward salt transport due to the tidal advective diffusion (line with triangles) and prescribed diffusion (line with asterisks). Gravitational circulation (solid line) and tidal rectification (dashed–dotted line) transport salt landward in the central region of salt intrusion. The stress-free surface condition (line with crosses) and tidal return flow (dashed line) transport salt landward only near the mouth. The total amount of salt transport is zero (dotted line in 2b) by definition, as the salt content in the estuary is constant in steady state.
2) Salt transport due to each process
In equilibrium, the divergence of the total (depth integrated) salt transport is zero at every location. However, the divergence of salt transport due to each physical transport process separately is not necessarily zero, indicating a local change of the salt content due to each process. If the divergence of the salt transport due to one process has locally a negative (positive) value, it implies this process contributes to an increase (a decrease) of salt content at that location.
Figures 3a–e show the divergence of the salt transport due to diffusive processes (including tidal advective diffusion and prescribed diffusion), gravitational circulation, tidal rectification, stress-free surface condition, and tidal return flow, respectively. Local convergences of salt are shown in dark gray, and divergences of salt are shown in white. The lines in Figs. 3a–e show the direction of salt transport due to each process. The solid (dashed–dotted) lines in Figs. 3b–e indicate that the salt is transported in a counterclockwise (clockwise) direction. Note that these lines are not the streamlines for the residual salt transport but for the residual water transport, since the divergence of the residual salt transport for each process individually does not vanish.
The divergence of the depth-integrated residual salt transport due to each process. Positive values indicate local decreases of salt content, while negative values indicate local increases of salt content. The lines show the directions of the depth-integrated residual salt transport due to each process, with counterclockwise salt transport shown by solid lines and clockwise transport by dashed–dotted lines in Figs. 3b–e.
Citation: Journal of Physical Oceanography 47, 7; 10.1175/JPO-D-16-0247.1
The diffusive processes transport salt from the shoals toward the deeper channel and subsequently toward the upstream (Fig. 3a). The diffusive processes result in an increase of salinity on the right side of the estuary and a decrease on the left. The convergences and divergences of the diffusive salt transport are significant, especially in the central region near the deeper channel, with a salt accumulation of up to 1 × 10−4 psu m s−1. Moreover, the salt accumulation on the right exceeds the salt decrease on the left side, resulting in an increase of salt content integrated over a cross section. This increase in salt content agrees with the decreasing longitudinal salt transport by diffusive processes, as shown in Fig. 2b.
The two-cell streamlines of gravitational circulation show that the density-driven flow transports salt landward into the estuary through the deeper channel, reverses direction in the central region, and then transports salt out of the estuary over the shoals (see Fig. 3b). The salt transport induced by gravitational circulation is asymmetric with respect to the midaxis due to the Coriolis deflection; the landward salt transport is deflected toward the left bank, and the seaward transport is deflected toward the right bank. Even though the longitudinal salt transport (integrated over a cross section) induced by gravitational circulation is much smaller than that induced by the diffusive processes, the gravitational circulation makes a significant contribution to local changes in salinity. The gravitational circulation induces a significant decrease of salt on the shoals, and a large increase of salt in the deeper channel, which reaches up to 5 × 10−5 psu m s−1 in the central region of salt intrusion.
The salt transport induced by residual flow due to tidal rectification, stress-free surface condition, and tidal return flow all show a similar pattern, with salt transported into the estuary over the shoals and out through the deeper channel (see Figs. 3c–e). The tidal rectification results in an accumulation of salt on the left side of the deeper channel, accompanied with a strong decrease of salt on the right. The stress-free surface condition and tidal return flow, however, contribute to a decrease of salt in the deeper channel and an accumulation of salt on the shoals.
b. Influence of estuarine geometry, Coriolis, and bathymetry
The salt transport processes in the default experiment are affected by the interplay between estuarine geometry, Coriolis deflection, and bathymetry. To investigate the influence of estuarine convergence on salt dynamics separately, an exponentially convergent channel with a horizontal bottom (no Coriolis) is studied in experiment I. In experiment II, the Coriolis force is added and hence the influence of the Coriolis force on salt dynamics can be found by comparing the results obtained in experiment II with those found in experiment I. In experiment III, the default bathymetry profile with a lateral channel–shoal structure [see Eq. (49)] is used, but now the Coriolis force is neglected. Hence, the influence of bathymetric variations on salt dynamics can be found by comparing the results from experiments III and I. The parameters used to define estuarine bathymetry and the Coriolis parameter are summarized in Table 2.
Model parameters for experiments I–III.
1) Influence of estuarine convergence (experiment I)
In a convergent channel without bathymetric variations or Coriolis force, the salt intrusion length is only 105 km, which is much smaller than that in the default experiment (120 km). The tidally averaged salinity is symmetric with respect to the midaxis, with the salinity near the banks slightly larger than that in the middle (see Fig. 4a). Since no circulation cells of salt transport are formed, the depth-integrated residual salt transport is zero at each location, and no streamlines are obtained.
As in Fig. 2, but for experiment I.
Citation: Journal of Physical Oceanography 47, 7; 10.1175/JPO-D-16-0247.1
The residual salt balance is identical to that obtained from the width-averaged model (Wei et al. 2016), where the seaward salt transport induced by river discharge is balanced by the landward transport by tidal advective diffusion and prescribed diffusion (Fig. 4b). The residual flow induced by gravitational circulation, tidal rectification, stress-free surface, and the tidal return flow, however, makes no contributions to the residual salt balance (integrated over the cross section).
The divergence of the residual salt transport induced by the diffusive processes, as shown in Fig. 5, indicates that diffusion results in a convergence (accumulation) of salt in the central region of the estuary. The salt transport paths induced by the diffusive processes are fully adapted to the estuarine geometry, with no horizontal circulations formed. It is important to note that even though no salt transport circulation cells are induced by including estuarine convergence, the salt dynamics can be significantly influenced by this factor. For instance, the estuarine convergence has a strong impact on the tidal water motion (Friedrichs and Aubrey 1994; Lanzoni and Seminara 1998) and river-induced flow due to changes of the cross-sectional area, thus affecting the salt transport contributions due to tidal advective diffusion (Wei et al. 2016) and river-induced flow (not shown).
As in Fig. 3a, but for the salt transport induced by diffusive processes in experiment I.
Citation: Journal of Physical Oceanography 47, 7; 10.1175/JPO-D-16-0247.1
2) Influence of Coriolis deflection (experiment II)
Including the Coriolis force in the convergent channel slightly increases the salt intrusion length to 108 km, with the salinity on the left larger than that on the right (see Fig. 6a). The salt transport pattern is significantly changed by the Coriolis deflection, with salt transported into the estuary from the left side of the channel and out from the right. Comparing the streamlines of the total salt transport in experiments I–II, it is found that the geometry-induced salt transport pattern as found in experiment I is replaced by a Coriolis-induced pattern in experiment II by including Coriolis force. It suggests that the influence of the Coriolis deflection on the total salt transport is much stronger than that of the estuarine convergence. Because of the Coriolis deflection, the seaward salt transport induced by river discharge is mainly balanced by the landward salt transport induced by diffusive processes, while the residual flow due to the stress-free surface condition and the tidal return flow makes a small but nonnegligible contribution (see Fig. 6b). The gravitational circulation, however, does not contribute to the residual salt balance.
As in Fig. 2, but for experiment II.
Citation: Journal of Physical Oceanography 47, 7; 10.1175/JPO-D-16-0247.1
Because of the Coriolis force, the landward salt transport by diffusive processes is deflected toward the left side of the estuary (see Fig. 7a), decreasing the cross-sectionally integrated longitudinal salt transport due to this contribution. It results in a significant divergence of salt on the left side and a convergence of salt on the right. The magnitude of local change in salinity induced by diffusive processes reaches up to 2 × 10−5 psu m s−1, much larger than that in experiment I where no Coriolis is included. The residual flow due to the stress-free surface condition and the tidal return flow tends to transport salt landward from the left side of the estuary and seaward from the right (see solid lines in Figs. 7b,c). It results in a remarkable accumulation of salt on the left side of the estuary and a divergence of salt on the right. The local change of salinity induced by the stress-free surface condition and the tidal return flow are up to 1 × 10−5 and 2 × 10−6 psu m s−1, respectively (see gray scales in Figs. 7b,c). The salt transport induced by tidal rectification and gravitational circulation are negligible (not shown).
As in Fig. 3, but for the salt transport induced by (a) diffusive processes, (b) stress-free surface condition, and (c) tidal return flow in experiment II.
Citation: Journal of Physical Oceanography 47, 7; 10.1175/JPO-D-16-0247.1
3) Influence of estuarine bathymetry (experiment III)
By including a lateral channel–shoal structure, the salt intrusion length is significantly increased to 117 km (see Fig. 8a), compared to experiment I. Furthermore, the tidally averaged salinities become higher in the deeper channel than those on the shoals. The residual salt transport patterns are also completely changed: salt is transported landward over the shoals and seaward through the deeper channel. This clearly shows the influence of bathymetric variations is larger than that of the estuarine convergence. By including the bathymetric variations, the landward salt transport contributions due to tidal advective diffusion, prescribed diffusion, and gravitational circulation are increased (see Fig. 8b). On the other hand, the residual flow induced by the stress-free surface condition and tidal return flow tends to transport salt seaward, in contrast to their landward salt transport contributions in experiment II. The salt transport due to tidal rectification is very small compared to the other processes.
As in Fig. 2, but for experiment III.
Citation: Journal of Physical Oceanography 47, 7; 10.1175/JPO-D-16-0247.1
Because of the lateral channel–shoal structure, the diffusive processes transport salt from the shoals to the deeper channel and then transport salt landward (see Fig. 9a). This is accompanied with an accumulation of salinity in the deeper channel and a decrease of salinity on the shoals. The gravitational circulation presents a landward flow in the deeper channel and a seaward flow on the shoals, as found by Wong (1994), transporting salt into the estuary through the deeper channel and out over the shoals (see Fig. 9b). It results in a strong convergence of salt in the deeper channel and a small divergence of salt on the shoals. The residual flow induced by tidal rectification, stress-free surface condition, and tidal return flow is characterized by a net landward flow over the shoals and balanced by a return (seaward) flow in the deeper channel (see Figs. 9c–e). The pattern of these three residual components is consistent with the topography-induced exchange flow, as found by Li and O’Donnell (1997) (without considering density), which tends to transport salt landward over the shoals and seaward through the deeper channel. This results in a divergence of salt in the deeper channel and an accumulation of salt on the shoals. The amount of local salt loss induced by the tidal return flow or stress-free surface condition is up to 5 × 10−5 psu m s−1. The salt change induced by the tidal rectification is very small (less than 1 × 10−5 psu m s−1) compared to the default experiment. This confirms that effects of earth rotation are very important in generating the tidally rectified flow, as reported in Huijts et al. (2009).
As in Fig. 3, but for experiment III.
Citation: Journal of Physical Oceanography 47, 7; 10.1175/JPO-D-16-0247.1
4. Discussion
The importance of various physical processes on salt intrusion length is evaluated by excluding each process individually and comparing corresponding changes in the salt intrusion length. It is found that the tidal advective diffusion is the predominant physical process for longitudinal salt intrusion, which is hardly influenced by residual flows (except for the river-induced flow). Since tidal advective diffusion is strongly influenced by many factors such as bottom friction, vertical mixing, external tidal forcing, estuarine bathymetry, and geometry (Wei et al. 2016), it is deduced that the salt intrusion length changes significantly with these factors. Though the residual circulations make no big contributions to salt intrusion lengths, they are crucial for the transverse structure of the tidally averaged salinity, which is essential for the transverse gravitational circulation and thence lateral sediment trapping in tidal estuaries (Huijts et al. 2006).
In section 4a, the salt transport due to the tidal advective diffusion is discussed: in section 4a(1), the salt transport induced by tidal advective diffusion in the idealized Delaware estuary is explained by comparing results of the experiments I–III with those of the default experiment; in section 4a(2), the resulting tidal advective diffusivity coefficients are discussed in relation to estuarine geometry, bathymetry, and Coriolis. The salt transport induced by the gravitational circulation, the strongest residual circulation process in the idealized Delaware estuary, is discussed in section 4b.
a. Tidal advective diffusion
1) Tidal advective salt flux
The transverse structure of the three-dimensional tidal advective salt flux (
Citation: Journal of Physical Oceanography 47, 7; 10.1175/JPO-D-16-0247.1
In experiment I, the longitudinal tidal advective salt flux is similar to that from a width-averaged model by Wei et al. (2016). The longitudinal tidal advective salt fluxes are nearly laterally uniform, with landward fluxes near the surface and seaward fluxes near the bottom (see the contours in Fig. 10a). Moreover, the estuarine convergence results in a lateral tidal current that interacts with the tidal salinity and transports salt toward the midaxis near the free surface and toward the banks near the bottom. These geometry-induced transverse tidal advective salt fluxes are smaller than 0.005 psu m s−1 (see arrows in Fig. 10a). When including the Coriolis effects (experiment II), |u0| and |S1| become larger on the left than the right (see Figs. 11a,b), while their phase differences remain almost unchanged (see Fig. 11c). This results in larger longitudinal tidal advective salt fluxes on the left than on the right (see Fig. 10b). In the transverse direction, the lateral tidal current induced by the Coriolis effects interacts with the tidal salinity and yields salt fluxes toward the left near the surface and toward the right near the bottom. This Coriolis-induced circulation, together with the geometry-induced circulation, results in a slightly enhanced counterclockwise salt transport on the right side of the estuary and a slightly weakened clockwise transport on the left (see arrows in Fig. 10b).
A qualitative view of the lateral profiles of |u0|, |S1|, and the phase difference ΦS − Φu influenced by (a)–(c) Coriolis effects and (d)–(f) lateral bathymetric variations. The amplitudes and phase difference near the surface are shown by solid lines, and those near the bottom are shown by dashed lines.
Citation: Journal of Physical Oceanography 47, 7; 10.1175/JPO-D-16-0247.1
In experiment III, tides propagate faster in the deeper channel than on the shoal, resulting in a larger |u0| and |S1| in the deeper channel than on the shoals (see Figs. 11d,e). Meanwhile, u0 in the deeper channel lags that on the shoals, and S1 on the shoals lags that in the deeper channel. Thus, in the deeper channel, the phase difference between u0 and S1 is smaller than 90° near the surface and slightly larger than 90° near the bottom; on the shoals, it is larger than 90° near the bottom, and smaller than 90° near the surface (see Fig. 11f). This results in enhanced landward tidal advective salt fluxes near the surface (relatively small seaward fluxes near the bottom) in the deeper channel. On the shoals, since |u0| and |S1| are much smaller than those in the deeper channel, the induced salt fluxes are much smaller (see gray scales in Fig. 10c). Integrated over the depth, the landward tidal advective salt transport is much stronger in the deeper channel. Moreover, the interaction between the tidal salinity and the lateral tidal current induced by the lateral depth variations results in axial-convergent tidal advective salt fluxes (up to 0.01 psu m s−1) toward the midaxis at all depths (see arrows in Fig. 10c).
Comparing the tidal advective salt fluxes in experiments I–III with the default experiment suggests that the lateral bathymetric variations dominate the longitudinal tidal advective salt fluxes in the idealized Delaware estuary. The lateral tidal advective salt fluxes, however, are influenced by estuarine convergence, Coriolis, and bathymetry. Because of the overall effects, the tidal advective salt fluxes toward the left on the top right of the channel as well as the fluxes toward the right on the bottom left are augmented (see Fig. 10d).
2) Tidal advective diffusivity
The strength of the (depth integrated) salt transport due to tidal advective diffusion is measured by the tidal advective diffusion coefficients, which result from the temporal correlation between the longitudinal/lateral tidal velocities and tidal salinities [see Eqs. (29)–(32)]. Since more processes are resolved in a three-dimensional model compared to a width-averaged approach, it is to be expected that the prescribed diffusion coefficient (including unresolved processes) has to be smaller compared to the values that have to be used in width-averaged models to qualitatively reproduce the observed salt intrusion in the Delaware estuary. Figure 12 shows the tidal advective diffusion coefficients for all experiments. Here, only 
The tidal advective diffusion coefficients for all experiments.
Citation: Journal of Physical Oceanography 47, 7; 10.1175/JPO-D-16-0247.1
The tidal advective diffusivity 
Coriolis effects significantly change the spatial pattern of 
b. Gravitational circulation
As found by Wei et al. (2016), the gravitational circulation does not influence the residual salt transport in convergent estuaries with a horizontal bed (with negligible lateral processes), since the near-bottom landward salt transport is canceled by the near-surface seaward transport (see gray scales in Figs. 13a,b). It suggests in estuaries with no depth variations, the gravitational circulation and salinity are hardly coupled, and the gravitational circulation can be understood straightforward using the obtained salinity.
(top) The residual water level induced by gravitational circulation (
Citation: Journal of Physical Oceanography 47, 7; 10.1175/JPO-D-16-0247.1
The estuarine convergence induces a lateral gradient of the tidally averaged salinity with S0 near the banks larger than that in the center. To balance this gradient, a barotropic pressure gradient is induced with a larger 
Because of the lateral bathymetric variations, the gravitational circulation transports salt landward through the deeper channel and seaward over the shoals (see gray scales in Figs. 13c,d). The longitudinal gravitational circulation is qualitatively consistent with Lerczak and Geyer (2004), with the strongest landward fluxes near the bottom of the deeper channel and the strongest seaward fluxes near the surface of the shoals. Nevertheless, there is a discrepancy: the gravitational circulation is generally landward near the bottom and seaward near the surface in Lerczak and Geyer (2004); in the present work, however, the landward flow is located at all depths of the deeper channel with the seaward flow on the shoals. This discrepancy is probably related to different vertical mixing coefficients used in two models; Lerczak and Geyer (2004) used a relatively small Aυ (0.002 m2 s−1) for their well-mixed runs, while in the present model Aυ is up to 0.005 m2 s−1 in the deeper channel. It is important to note that in the presence of the channel–shoal structure, the gravitational circulation and the salinity distribution become strongly coupled, and hence they cannot be interpreted straightforwardly, as in experiments I–II.
In experiment III, the diffusive processes and the residual flows act together, resulting in a larger (smaller) S0 (
5. Response to river discharge
To systematically explore the response of salt intrusion to river discharge, the model is applied to experiment III using three different river discharge values: 72, 288, and 864 m3 s−1, respectively. It is found that the simulated, tidally averaged, salt intrusion length varies significantly from 120 km (for Q = 72 m3 s−1) to 80 km (Q = 288 m3 s−1) and 60 km (Q = 864 m3 s−1). This strong dependency of salt intrusion length on river discharge is not consistent with observations of Garvine et al. (1992), who found the salt intrusion length only weakly responds to the increase of river discharge in the Delaware estuary. Earlier studies have shown that increasing river discharge can increase stratification and lead to weaker vertical mixing, consequently enhancing the residual circulation (Monismith et al. 2002; Ralston et al. 2008). This enhanced residual circulation is found to play an important role in transporting salt landward and weakening the dependency of salt intrusion on the variations of river discharge (Ralston et al. 2008; Lerczak et al. 2009). Therefore, to understand the overestimated sensitivity of the simulated salt intrusion to river discharge, the stratification (measured by the top-to-bottom salinity difference) at each location is calculated for different river discharges (not shown). It is found that the stratification in the central region of salt intrusion increases with increasing river discharge, consistent with model studies (Ralston et al. 2008; Lerczak et al. 2009) and observations (McSweeney et al. 2016). The top-to-bottom salinity differences (less than 1.5 psu), however, are much smaller than the observed values during large river discharges (McSweeney et al. 2016). This implies that the relatively strong stratification in case of large river discharges is not resolved in this model, which is probably the reason for the overestimated impact of river discharge on salt intrusion. Therefore, to reproduce the observed salt intrusion for large river discharges using the model presented in this paper, a larger horizontal eddy diffusivity coefficient needs to be prescribed. This is consistent with Gay and O’Donnell (2009), who captured the weak response of the salt intrusion length to river discharge in the Delaware estuary by allowing the along-channel dispersion coefficient to exponentially increase with river discharge.
6. Conclusions
A semianalytical, three-dimensional model is developed to investigate the three-dimensional salt dynamics in idealized well-mixed estuaries. The adopted perturbation method allows for a systematic decomposition of the total salt transport into contributions due to various physical processes. Each salt transport process and its relative importance can be investigated in isolation.
First, the salt dynamics for a schematized estuary is investigated. The schematized estuary is exponentially converging, and the bathymetry shows a channel–shoal structure. In this experiment, the tidal advective diffusion, induced by the temporal correlation between horizontal tidal velocities and salinity (including the lateral processes), is the dominant landward salt transport process. The cross-sectionally integrated salt transport induced by residual circulations due to gravitational circulation, tidal rectification, stress-free surface condition, and tidal return flow is small compared to that induced by the tidal advective diffusion. The salt transport due to these residual circulations, however, is locally significant. The gravitational circulation transports salt landward through the deeper channel and seaward over the shoals. The residual circulations due to other mechanisms are smaller and transport salt in the opposite direction compared to the transport pattern of gravitational circulation.
The residual salt balance is influenced by the estuarine convergence, bathymetry, and Coriolis. Three experiments are conducted in order to investigate these three factors on the estuarine salt transport in separation: in experiment I, an exponentially convergent estuary is used with a horizontal bed, no Coriolis; in experiment II, the estuarine geometry and bathymetry are the same as in experiment I, but the Coriolis force is included; and in experiment III, a channel–shoal structure is included while the Coriolis effects are excluded. The tidal advective diffusion is found to be the most dominant landward salt transport process for all experiments. The salt intrusion length slightly increases, compared to the results of experiment I, when Coriolis is included (experiment II) and increases significantly by including the channel–shoal structure (experiment III). The contribution of tidal advective diffusion to estuarine salt transport can be measured by the tidal advective diffusion coefficients, which can be explicitly calculated and are strongly dependent on the Coriolis effects and lateral bathymetric variations.
The residual circulations hardly influence the salt intrusion length, but are crucial for the local salt transport. In experiment I, the depth-integrated residual circulations vanish, with the river-induced seaward salt transport balanced by diffusive processes only. No significant lateral salt transport patterns are induced by the width convergence of the estuary. In experiment II (including Coriolis deflection), the residual circulations due to tidal rectification and the stress-free surface condition transport salt into the estuary from the left side of the estuary (looking seaward) and out on the right side, resulting in a significant contribution to the landward salt transport. In experiment III, salt transport due to all processes becomes stronger due to the channel–shoal structure. The gravitational circulation, as the most significant residual circulation process, is characterized by a symmetric two-cell circulation pattern. Residual circulations due to other processes are relatively weak. The residual salt transport due to residual circulations is essential for the lateral salinity structure, which can drive a strong transverse gravitational circulation within the salt intrusion region.
The authors appreciate the suggestion from Dr. George P. Schramkowski (Flanders Hydraulics Research) about visualizing salt transport patterns using streamlines. This project is mainly funded by the China Scholarship Council (CSC, File 201206710049). This work was sponsored by NWO Exacte Wetenschappen (Physical Sciences) for the use of supercomputer facilities, with financial support from the Nederlandse Organisatie voor Wetenschappelijk Onderzoek [Netherlands Organization for Scientific Research (NWO)]. The work is part of the research programme NWO–ALW project 8843.10.005, which is financed by the NWO and the Chinese Organisation for Scientific Research (NSFC).
APPENDIX A
Scaling Analysis
A perturbation method is used to analytically solve the water motion and salinity. First of all, variables are scaled with their typical scales (see Table A1), denoted by a tilde (~). This model does not resolve the boundary layer, and processes in typical estuarine length scales are focused. The variations of water motion and salinity in the longitudinal and lateral directions are assumed to scale with the typical estuarine length scale Lb. For estuaries with an exponentially decreasing width, Lb is equivalent to the estuarine convergence length [see Eq. (46)]. The variations in the vertical are assumed to scale with the estuarine water depth.
Scales of physical variables.
As a next step, the magnitudes of the scaling parameters are compared with a small parameter 
Order of magnitude of scaling parameters.
Then, we substitute the magnitudes of these scaling parameters into Eqs. (A2)–(A9), which yields the dimensionless governing equations and boundary conditions for water motion and salinity in terms of ε. Substituting the expansions of each dimensionless variable in terms of ε and collecting the terms at the same order of ε, we derived the dimensional governing equations and boundary conditions at different order of ε, which will be used to calculate the water motion and salinity semianalytically.
APPENDIX B
Leading-Order Water Motion
The leading-order water motion is forced by the semidiurnal tide; thus, it consists of only the M2 tidal constituent.
APPENDIX C
First-Order Salinity Equations
APPENDIX D
First-Order Residual Flow
a. Governing equations
b. Gravitational circulation
c. The seaward boundary condition
When solving this coupled system of equations, special care has to be taken of the seaward boundary, since numerical instabilities can occur when using an arbitrary prescribed water level at this boundary (Chen and Sanford 2009). This is because an inconsistent prescribed sea surface elevation at the mouth can result in a rapid adjustment to a consistent solution in the interior domain. This rapid adjustment will result in large unphysical gradients in the water motion close to the boundary and thus cause numerically inaccurate solutions.
To avoid that, the computational domain is extended by 30 km in the seaward direction. At the open boundary of the computational domain, a constant M2 tidal water level is prescribed, and the tidal water motion is calculated. This results in a physically consistent water motion within the domain of interest. Then, the resulting M2 tidal water elevation at the seaward boundary of the physical domain is rescaled, such that the width-averaged tidal amplitude at the original open boundary (now located at x = 30 km) is the same as the tidal amplitude that is prescribed at the open boundary of the physical domain. This will result in a consistent M2 water motion in the domain of interest. The extended domain is then used to calculate the residual flow using the calculated tidal water motion, except for the residual flow generated by the baroclinic pressure. To satisfy the physical boundary, a constant residual water level 
REFERENCES
Aristizábal, M., and R. Chant, 2013: A numerical study of salt fluxes in Delaware Bay estuary. J. Phys. Oceanogr., 43, 1572–1588, doi:10.1175/JPO-D-12-0124.1.
Becherer, J., M. T. Stacey, L. Umlauf, and H. Burchard, 2015: Lateral circulation generates flood tide stratification and estuarine exchange flow in a curved tidal inlet. J. Phys. Oceanogr., 45, 638–656, doi:10.1175/JPO-D-14-0001.1.
Buijsman, M., and H. Ridderinkhof, 2008: Variability of secondary currents in a weakly stratified tidal inlet with low curvature. Cont. Shelf Res., 28, 1711–1723, doi:10.1016/j.csr.2008.04.001.
Burchard, H., R. D. Hetland, E. Schulz, and H. M. Schuttelaars, 2011: Drivers of residual estuarine circulation in tidally energetic estuaries: Straight and irrotational channels with parabolic cross section. J. Phys. Oceanogr., 41, 548–570, doi:10.1175/2010JPO4453.1.
Chen, S.-N., and L. P. Sanford, 2009: Lateral circulation driven by boundary mixing and the associated transport of sediments in idealized partially mixed estuaries. Cont. Shelf Res., 29, 101–118, doi:10.1016/j.csr.2008.01.001.
Chen, S.-N., W. R. Geyer, D. K. Ralston, and J. A. Lerczak, 2012: Estuarine exchange flow quantified with isohaline coordinates: Contrasting long and short estuaries. J. Phys. Oceanogr., 42, 748–763, doi:10.1175/JPO-D-11-086.1.
Cushman-Roisin, B., and J.-M. Beckers, 2011: Introduction to Geophysical Fluid Dynamics: Physical and Numerical Aspects. International Geophysics Series, Vol. 101, Academic Press, 875 pp.
Dyer, K., 1974: The salt balance in stratified estuaries. Estuarine Coastal Mar. Sci., 2, 273–281, doi:10.1016/0302-3524(74)90017-6.
Fischer, H., 1972: Mass transport mechanisms in partially stratified estuaries. J. Fluid Mech., 53, 671–687, doi:10.1017/S0022112072000412.
Friedrichs, C. T., and D. G. Aubrey, 1994: Tidal propagation in strongly convergent channels. J. Geophys. Res., 99, 3321–3336, doi:10.1029/93JC03219.
Friedrichs, C. T., and J. M. Hamrick, 1996: Effects of channel geometry on cross sectional variations in along channel velocity in partially stratified estuaries. Buoyancy Effects on Coastal and Estuarine Dynamics, D. G. Aubrey and C. T. Friedrichs, Eds., Wiley, 283–300.
Garvine, R. W., R. K. McCarthy, and K.-C. Wong, 1992: The axial salinity distribution in the Delaware estuary and its weak response to river discharge. Estuarine Coastal Shelf Sci., 35, 157–165, doi:10.1016/S0272-7714(05)80110-6.
Gay, P., and J. O’Donnell, 2009: Buffering of the salinity intrusion in estuaries by channel convergence. Hydrol. Earth Syst. Sci. Discuss., 6, 6007–6033, doi:10.5194/hessd-6-6007-2009.
Geyer, W. R., and P. MacCready, 2014: The estuarine circulation. Annu. Rev. Fluid Mech., 46, 175–197, doi:10.1146/annurev-fluid-010313-141302.
Gockenbach, M. S., 2006: Understanding and Implementing the Finite Element Method. SIAM, 363 pp.
Guymer, I., and J. West, 1992: Longitudinal dispersion coefficients in estuary. J. Hydraul. Eng., 118, 718–734, doi:10.1061/(ASCE)0733-9429(1992)118:5(718).
Hansen, D. V., and M. Rattray, 1965: Gravitational circulation in straits and estuaries. Estuarine Coastal Mar. Sci., 23, 104–122.
Hughes, F., and M. Rattray, 1980: Salt flux and mixing in the Columbia River estuary. Estuarine Coastal Mar. Sci., 10, 479–493, doi:10.1016/S0302-3524(80)80070-3.
Huijts, K., H. Schuttelaars, H. De Swart, and A. Valle-Levinson, 2006: Lateral entrapment of sediment in tidal estuaries: An idealized model study. J. Geophys. Res., 111, C12016, doi:10.1029/2006JC003615.
Huijts, K., H. Schuttelaars, H. De Swart, and C. Friedrichs, 2009: Analytical study of the transverse distribution of along-channel and transverse residual flows in tidal estuaries. Cont. Shelf Res., 29, 89–100, doi:10.1016/j.csr.2007.09.007.
Huijts, K., H. E. De Swart, G. P. Schramkowski, and H. M. Schuttelaars, 2011: Transverse structure of tidal and residual flow and sediment concentration in estuaries. Ocean Dyn., 61, 1067–1091, doi:10.1007/s10236-011-0414-7.
Jay, D., and J. Smith, 1990: Residual circulation in shallow estuaries: 2. Weakly-stratified and partially mixed, narrow estuaries. J. Geophys. Res., 95, 733–748, doi:10.1029/JC095iC01p00733.
Kuijper, K., and L. C. Van Rijn, 2011: Analytical and numerical analysis of tides and salinities in estuaries; Part II: Salinity distributions in prismatic and convergent tidal channels. Ocean Dyn., 61, 1743–1765, doi:10.1007/s10236-011-0454-z.
Kumar, M., H. M. Schuttelaars, P. C. Roos, and M. Möller, 2016: Three-dimensional semi-idealized model for tidal motion in tidal estuaries. Ocean Dyn., 66, 99–118, doi:10.1007/s10236-015-0903-1.
Kumar, M., H. M. Schuttelaars, and P. C. Roos, 2017: Three-dimensional semi-idealized model for estuarine turbidity maxima in tidally dominated estuaries. Ocean Modell., 113, 1–21, doi:10.1016/j.ocemod.2017.03.005.
Lacy, J. R., and S. G. Monismith, 2001: Secondary currents in a curved, stratified, estuarine channel. J. Geophys. Res., 106, 31 283–31 302, doi:10.1029/2000JC000606.
Lanzoni, S., and G. Seminara, 1998: On tide propagation in convergent estuaries. J. Geophys. Res., 103, 30 793–30 812, doi:10.1029/1998JC900015.
Lerczak, J. A., and W. R. Geyer, 2004: Modeling the lateral circulation in straight, stratified estuaries. J. Phys. Oceanogr., 34, 1410–1428, doi:10.1175/1520-0485(2004)034<1410:MTLCIS>2.0.CO;2.
Lerczak, J. A., W. R. Geyer, and R. J. Chant, 2006: Mechanisms driving the time-dependent salt flux in a partially stratified estuary. J. Phys. Oceanogr., 36, 2296–2311, doi:10.1175/JPO2959.1.
Lerczak, J. A., W. R. Geyer, and D. K. Ralston, 2009: The temporal response of the length of a partially stratified estuary to changes in river flow and tidal amplitude. J. Phys. Oceanogr., 39, 915–933, doi:10.1175/2008JPO3933.1.
Li, C., and J. O’Donnell, 1997: Tidally driven residual circulation in shallow estuaries with lateral depth variation. J. Geophys. Res., 102, 27 915–27 929, doi:10.1029/97JC02330.
MacCready, P., 2004: Toward a unified theory of tidally-averaged estuarine salinity structure. Estuaries, 27, 561–570, doi:10.1007/BF02907644.
MacCready, P., and W. R. Geyer, 2010: Advances in estuarine physics. Annu. Rev. Mar. Sci., 2, 35–58, doi:10.1146/annurev-marine-120308-081015.
McCarthy, R. K., 1993: Residual currents in tidally dominated, well-mixed estuaries. Tellus, 45A, 325–340, doi:10.3402/tellusa.v45i4.14896.
McSweeney, J. M., R. J. Chant, and C. K. Sommerfield, 2016: Lateral variability of sediment transport in the Delaware estuary. J. Geophys. Res. Oceans, 121, 725–744, doi:10.1002/2015JC010974.
Monismith, S. G., W. Kimmerer, J. R. Burau, and M. T. Stacey, 2002: Structure and flow-induced variability of the subtidal salinity field in northern San Francisco Bay. J. Phys. Oceanogr., 32, 3003–3019, doi:10.1175/1520-0485(2002)032<3003:SAFIVO>2.0.CO;2.
Nunes, R., and J. Simpson, 1985: Axial convergence in a well-mixed estuary. Estuarine Coastal Shelf Sci., 20, 637–649, doi:10.1016/0272-7714(85)90112-X.
Oey, L.-Y., G. L. Mellor, and R. I. Hires, 1985: A three-dimensional simulation of the Hudson–Raritan estuary. Part I: Description of the model and model simulations. J. Phys. Oceanogr., 15, 1676–1692, doi:10.1175/1520-0485(1985)015<1676:ATDSOT>2.0.CO;2.
Ralston, D. K., and M. T. Stacey, 2005: Longitudinal dispersion and lateral circulation in the intertidal zone. J. Geophys. Res., 110, C07015, doi:10.1029/2005JC002888.
Ralston, D. K., W. R. Geyer, and J. A. Lerczak, 2008: Subtidal salinity and velocity in the Hudson River estuary: Observations and modeling. J. Phys. Oceanogr., 38, 753–770, doi:10.1175/2007JPO3808.1.
Ralston, D. K., W. R. Geyer, and J. A. Lerczak, 2010: Structure, variability, and salt flux in a strongly forced salt wedge estuary. J. Geophys. Res., 115, C06005, doi:10.1029/2009JC005806.
Schramkowski, G. P., and H. De Swart, 2002: Morphodynamic equilibrium in straight tidal channels: Combined effects of Coriolis force and external overtides. J. Geophys. Res., 107, 3227, doi:10.1029/2000JC000693.
Schulz, E., H. M. Schuttelaars, U. Gräwe, and H. Burchard, 2015: Impact of the depth-to-width ratio of periodically stratified tidal channels on the estuarine circulation. J. Phys. Oceanogr., 45, 2048–2069, doi:10.1175/JPO-D-14-0084.1.
Smith, R., 1977: Long-term dispersion of contaminants in small estuaries. J. Fluid Mech., 82, 129–146, doi:10.1017/S0022112077000561.
Valle-Levinson, A., Ed., 2010: Contemporary Issues in Estuarine Physics. Cambridge University Press, 315 pp.
Valle-Levinson, A., K.-C. Wong, and K. M. Lwiza, 2000: Fortnightly variability in the transverse dynamics of a coastal plain estuary. J. Geophys. Res., 105, 3413–3424, doi:10.1029/1999JC900307.
Valle-Levinson, A., C. Reyes, and R. Sanay, 2003: Effects of bathymetry, friction, and rotation on estuary–ocean exchange. J. Phys. Oceanogr., 33, 2375–2393, doi:10.1175/1520-0485(2003)033<2375:EOBFAR>2.0.CO;2.
Wang, T., W. R. Geyer, P. Engel, W. Jiang, and S. Feng, 2015: Mechanisms of tidal oscillatory salt transport in a partially stratified estuary. J. Phys. Oceanogr., 45, 2773–2789, doi:10.1175/JPO-D-15-0031.1.
Wei, X., G. P. Schramkowski, and H. M. Schuttelaars, 2016: Salt dynamics in well-mixed estuaries: Importance of advection by tides. J. Phys. Oceanogr., 46, 1457–1475, doi:10.1175/JPO-D-15-0045.1.
West, J., R. Uncles, J. Stephens, and K. Shiono, 1990: Longitudinal dispersion processes in the upper tamar estuary. Estuaries, 13, 118–124, doi:10.2307/1351580.
Winant, C. D., 2008: Three-dimensional residual tidal circulation in an elongated, rotating basin. J. Phys. Oceanogr., 38, 1278–1295, doi:10.1175/2007JPO3819.1.
Wong, K.-C., 1994: On the nature of transverse variability in a coastal plain estuary. J. Geophys. Res., 99, 14 209–14 222, doi:10.1029/94JC00861.
