Estuarine Plume

Download as pdf or txt
Download as pdf or txt
You are on page 1of 14

Continental Shelf Research 116 (2016) 122–135

Contents lists available at ScienceDirect

Continental Shelf Research


journal homepage: www.elsevier.com/locate/csr

Research papers

Sediment transport in a surface-advected estuarine plume


H.Y. Yao a,b,n, N. Leonardi b,c, J.F. Li a, S. Fagherazzi b
a
State Key Laboratory of Estuarine and Coastal Research‚ East China Normal University, Shanghai 200062, China
b
Department of Earth and Environment, Boston University, 675 Commonwealth Avenue, Office STO 141G, Boston, MA 02215, USA
c
Department of Geography and Planning, University of Liverpool, Liverpool L69 7ZT, UK

art ic l e i nf o a b s t r a c t

Article history: The interplay between suspended-sediment transport and plume hydrodynamics in a surface-advected
Received 13 February 2015 estuarine plume is studied using a three-dimensional numerical model. Our analysis focuses on the
Received in revised form formation of a sediment-rich alongshore current and on the effect of sediments on the structure of the
17 January 2016
recirculating freshwater bulge. We introduce the ratio Y between the traveling time of sediment along
Accepted 23 January 2016
the bulge edge and the settling timescale. When Y o1, suspended sediments enter the alongshore
Available online 26 January 2016
coastal current. When Y 41 the sediments are deposited within the bulge. We find that a critical range of
Keywords: settling velocities exist above which no transport in the costal current is allowed. Critical settling-velocity
Estuarine plume values increase with river discharge. Therefore, low magnitude and long-lasting floods promote sedi-
Sediment transport
ment sorting in the continental shelf. We further find that, for a given flood duration, intermediate flood
Coastal current
magnitudes at the limit between subcritical and supercritical flow maximize the alongshore sediment
Freshwater transport
Flood transport. Similarly, for a fixed input of water and sediments, intermediate discharge durations maximize
alongshore sediment transport.
& 2016 Elsevier Ltd. All rights reserved.

1. Introduction Numerical results using a semi-analytical 2.5D model indicate that


in marine environments where the depth of water exceeds the
Buoyant fresh water discharges are one of the main drivers for Ekman scale (typically 10–30 m), the direction of suspended se-
coastal and shelf currents and their hydrodynamics determine the diment transport in the lower water column is different from the
fate and transport of sediments, nutrients and pollutants entering depth integrated water flow direction because of velocity veering
the ocean (Geyer et al., 2004a; Liu et al., 2007; Lohrenz et al., 1999; induced by the Coriolis effect (Shapiro, 2004).
Warne et al., 2002; Wright, 1977). Sediment plumes are thus Numerical (e.g. Chao and Boicourt, 1986; Marsaleix et al.,1998),
crucial for the morphodynamic evolution of coastal areas, and and empirical (e.g. Masse and Murthy, 1992) investigations con-
understanding their complex dynamics would promote a more firm that the river plume can be divided into two main regions: a
effective management of these environments. baroclinic eddy that forms near the river mouth (the recirculating
Many buoyant inflows remain trapped at the surface of shelf bulge) and an alongshore coastal current (Fig. 1). Among others,
water, forming a thin plume on top of ambient water and main- noticeable examples of anti-cyclonic rotating bulges are those
discharged from Chesapeake Bay and Delaware Bay in the USA and
taining little or no contact with the shelf bottom while spreading
the Yangtze River in China (Geyer et al., 2004b; Yu, 2006). Fresh-
offshore (Geyer et al., 2004b; Yankovsky and Chapman, 1997). A
water bulge grows in time and as long as its radius is relatively low
well-documented feature of surface-trapped river flows is their
it remains attached to the coast with a large freshwater fraction
deflection by the Coriolis force which triggers a flow circulation
delivered to the coastal current. When the bulge size exceeds a
within the plume (Geyer et al., 2004b; Yu, 2006). Near the river
critical limit, it separates from the coast and cyclically regrows,
mouth, flow inertia can be considered the main hydrodynamic causing the flow in the coastal current to be intermittent (Horner-
driver. Farther from it, earth rotation causes the river outflow to Devine et al., 2006). The higher is the discharge the bigger is the
evolve into a coastal current flowing in the direction of a propa- freshwater bulge and the faster the bulge becomes unstable. For
gating Kelvin wave (i.e. with the shore on the right in the Northern supercritical flow (densimetric Froude number higher than one)
hemisphere, see (Geyer et al., 2004b; Nof and Pichevin, 2001). the bulge becomes unstable after 5 rotations (e.g. after around
5 days, Horner-Devine et al., 2006). Fig. 2 shows example of
n
Corresponding author at: State Key Laboratory of Estuarine and Coastal Re-
freshwater bulge at the end of the 5th numerical day and for dif-
search East China Normal University, Shanghai 200062, China ferent discharge conditions; in panel D we show an example of
E-mail address: jason.a.yao@hotmail.com (H.Y. Yao). unstable bulge starting to detach from the coast: the location

http://dx.doi.org/10.1016/j.csr.2016.01.014
0278-4343/& 2016 Elsevier Ltd. All rights reserved.
H.Y. Yao et al. / Continental Shelf Research 116 (2016) 122–135 123

List of parameters S (x , y ) Surface salinity at (x, y ), Psu


So Salinity of ambient sea water, Psu
Parameters Meaning, Units Sr River salinity, Psu
B River mouth width, M ∆S Salinity anomaly between river discharge and ambient
C Suspended sediment concentration in the plume, sea water, psu
kg/m3 Tst Sediment transport time scale, s
Cr Concentration of suspended sediment, kg/m3 Tωs Sediment settling time scale, s
C (x, y ) Surface suspended sediment concentration at (x, y ), v Flow velocity, m/s
kg/m3 vr River flow velocity, m/s
f Coriolis parameter, s  1 xf Horizontal distance from freshwater center to shore-
G′ Reduced gravitational acceleration with sediment, m2 line, m
/s Y Sediment time scale ratio
g Gravitational acceleration, m2/s yf Longitudinal distance from freshwater center to ex-
g′ Reduced gravitational acceleration without sediment, tension cord of mouth, m
m2/s xi Horizontal distance from calculated point to shoreline,
H River mouth depth, M m
h Plume depth, M yi Longitudinal distance from calculated point to exten-
ho Plume depth at the coast, M sion cord of mouth, m
h̅ Average plume depth, M α Constant coefficient
L Plume width, M β Constant coefficient
Li Inertial radius, M γ Constant coefficient
m Constant coefficient ρo Density of ambient sea water, kg/m3
Q River discharge, m3/s ρr River discharge density without sediment, kg/m3
Q fcc Freshwater transport in the coastal current with se- ρs Density of suspended sediment, kg/m3
diment, m3/s ∆ρ Density anomaly between river fluid and ambient sea
Q fcc0 Freshwater transport in the coastal current without water without sediment, kg/m3
sediment, m3/s ∆ρ′ Density anomaly between plume and ambient sea
∆Q fcc Decrease of freshwater transport in the coastal cur- water with sediment, kg/m3
rent, m3/s θf Angle between freshwater center-mouth and shore-
Q ssc Sediment transport in the coastal current, kg/s line, °
Rf Distance from freshwater center to mouth, M θc Angle between sediment distributional center-mouth
Rc Distance from sediment distributional center to and shoreline, °
mouth, M ωs Settling velocity of suspended sediment, mm/s
rs Bulge radius, M ωcs Critical settling velocity, mm/s
S Salinity at calculated point, Psu

where the outer bulge current impinges on the wall has moved 1988). The relationship between suspended sediment concentra-
toward the inlet, cutting off the flow of the alongshore current. tion and settling velocity has been explored by many researchers
Numerous studies have been carried out to gain physical in- (Gibbs, 1985; Krone, 1962; Winterwerp, 2002; You, 2004) and it
sight into the dynamics of surface trapped river plumes using ei- was found that settling velocity is independent of suspended se-
ther laboratory experiments (Avicola and Huq, 2003a, b; Horner- diment concentration C when C o0.3 kg/m3, while it increases as a
Devine et al., 2006), field measurements (Hickey et al., 1998; function of concentration for 0.3 o Co 4.3 kg/m3 due to floccula-
Masse and Murthy, 1992), and analytical and numerical models tion (You, 2004). However, flocculation likely affects only a part of
(Beardsley and Hart, 1978; Chao and Boicourt, 1986; Fong and the sediment load, and a fraction of very fine sediments can be still
Geyer, 2002; Nof and Pichevin, 2001). However, these studies transported far from the river mouth. For example, suspended
mainly focused on the structure and dynamics of the freshwater sediment concentrations in the turbidity maximum zone of the
bulge and of the coastal current. Less attention has been paid to Yangtze Estuary varies from 2 to 10 kg/m3, with flocculation
the effect of Coriolis forces on sediment dispersal and transport trapping large amounts of sediment inside the river mouth (Li and
(see Hetland and Hsu, 2013). Zhang, 1998). Based on calculations performed by Milliman (1985)
The aim of this study is to extend previous work on the hy- and Liu (2007), it is believed that about 30% of total sediment flux
drodynamic of surface trapped river plumes to systems where fine discharged from the Yangtze River is transported several hundred
sediments are also present. Specifically, we use numerical ex- kilometers southward and deposits along the Zhejiang–Fujian
periments to analyze the reciprocal interactions between sedi- coastal zone. Our results are mainly relevant for the sediment
ment transport and plume hydrodynamics. We focus on the effect fraction that is transported far from the river mouth and con-
of sediments on the structure of the recirculating freshwater bulge tributes to the along-shelf sediment diffusion (Driscoll and Karner,
and coastal current and on the effect of centrifugal forces on se- 1999). The manuscript is organized as follow: after describing the
diment delivery to the ocean. Special attention is given to the numerical model and numerical-model setup, we define a series of
study of the alongshore sediment transport in the coastal current. variables and parameters used in the investigation. The results
We further focus on fine cohesive sediments that do not deposit in section is divided into two main parts: in the first part we focus on
proximity of the river mouth and are more likely transported the effect of sediments on the structure and geometry of the
within the freshwater bulge and in the alongshore current. freshwater bulge and freshwater transport in the alongshore
For fine cohesive sediments, the settling velocity also depends coastal currents. This section explores the effect of changes in
on flocculation at high sediment concentrations (Van Leussen, water density caused by sediments, in analogy with salinity
124 H.Y. Yao et al. / Continental Shelf Research 116 (2016) 122–135

Fig. 1. Domain of the numerical model and computational mesh used in the simulations.

induced density anomalies. In the second part we focus on the reasonable computational time. For all simulations, a river dis-
alongshore sediment transport. We identify optimum sediment charge was prescribed to enter at y ¼302 km (Fig. 1). A zero flux
input concentrations and sediment settling velocities maximizing boundary condition was used for the upper and lower boundaries,
the alongshore sediment transport. We further determine critical and a constant water level was imposed at the right boundary.
discharge values and flood durations that enhance the alongs- The width and depth of the river mouth are 3 km and 15 m
hore sediment transport. Finally, we compare our results to real respectively, and they are based on realistic riverine systems. For
systems. example, the width and depth of the main navigation channel of
the Yangtze River, North Passage, are around 3 km and 15 m. The
salinity of the input water is set to 28 psu in accordance with
2. Model description (Fong and Geyer, 2002), while the ambient water salinity is 32 psu.
Density anomaly is thus set around 3 kg/m3, which is a typical
We use the numerical model Delft3D (Lesser et al., 2004) that value for estuarine inflows (Yankovsky et al., 2001).
solves the Navier–Stokes equations for turbulent flow and sedi- The sediment input consists of a cohesive sediment fraction
ment transport. These equations consist of the horizontal mo- (mud), with settling velocity ranging from 10  3 to 5 mm/s. These
mentum equation, the continuity equation, the sediment transport low settling velocities where chosen to study sediment transport
equation, and a turbulence closure module. In this model the within the plume area, since the sandy sediment fraction of a river
vertical momentum equation reduces to the hydrostatic pressure typically settles inside or near the river mouth (see Fagherazzi
relationship because vertical accelerations are considered negli- et al., 2015 for a review). Different concentration values are used,
gible with respect to gravitational ones (Lesser et al., 2004). varying from 0.1 to 4 kg/m3 while the discharge values at the
Model geometry is showed in Fig. 1. The domain size is mouth range from 2500 to 30,000 m3/s. Specifically, very high
110  360 km2, the grid cells are 3  3 km2 near the lower discharge and sediment concentration values have been used to
boundary and they are refined to 1  3 km2 near the river mouth. simulate large flood conditions. This is in agreement with em-
To gain insight into the vertical structure of the river plume we use pirical relationships between in-channel water discharge and
20 vertical layers. Sensitivity tests were performed with a number suspended sediment concentration (C–Q relations), indicating that
of vertical layers between 20 and 40 and the results based on 20 sediment concentration is commonly increasing with discharge in
2layers were found to be a good compromise between computa- rivers (Williams, 1989).
tional time and model accuracy. The upper 12 layers have 1.5 m The model uses the k-ε closure model to determine vertical
thickness, while the lower 8 layers have a 4 m thickness. The time eddy viscosity and diffusivity. The Coriolis parameter is set to
step was set to 2 min to guarantee model stability and a 7.5  10  5/s. This value corresponds to 30° north latitude and
H.Y. Yao et al. / Continental Shelf Research 116 (2016) 122–135 125

Fig. 2. Freshwater bulge represented by surface salinity at t ¼ 5 days for: (A) Q ¼ 10,000 m3/s, (B) Q ¼ 20,000 m3/s, (C) Q ¼ 30,000 m3/s, (D) Q ¼ 50,000 m3/s. In panel C the
center of gravity of the freshwater bulge is also indicated. Notice that in panel D the system is unstable, and the freshwater bulge is detaching from the wall.

could thus be representative of some of the largest riverine sys- the planar coordinates of the center of gravity of the freshwater
tems in the world, such as the Yangtze River, China, the Mississippi bulge, xc and yc are the planar coordinates of the center of gravity
River, USA, or the Nile River, Egypt. Since we are mainly focused on of the sediment bulge, Rf and Rc are the distance from the river
the reciprocal interactions between sediment transport processes mouth of the center of gravity of the freshwater bulge and
and river plume structure, simulations are not intended to discuss of the sediment bulge respectively. θf , and θc are the angles of the
the morphologic evolution of the shelf bottom. Topographical freshwater and sediment bulge with respect to the shoreline
update is therefore prevented and the effect of wind waves (e.g. (Fig. 2).
Nardin and Fagherazzi, 2012, Nardin et al., 2013) and tides (e.g. The freshwater transport into the alongshore coastal current
Leonardi et al. 2013, 2015) on sediment delivery is neglected as can be schematized as a simple two-layer Margules front (Fong
well. and Geyer, 2002; Garvine, 1999; Lentz and Helfrich, 2002; Pimenta
et al., 2011). In a Margules front system, assuming quiescent lower
layers and a geostrophic cross shore momentum, the alongshore
velocity can be written as (Fong and Geyer, 2002):
3. Model parameters
∆ρ ∂h
v=g (3)
The salinity anomaly s is used to mark the passage from the ρ0 f ∂x
denser ambient water to the freshwater plume (Garvine, 1999):
where ∆ρ is the density anomaly between ambient sea and plume
So − S water without sediment, h is the plume depth, ρ0 is the density of
s=
So − Sr (1) ambient sea water, f is the Coriolis parameter and g is the grav-
itational acceleration.
where So is ambient water salinity, Sr is the river discharge salinity,
In the absence of any sediment input from the river mouth,
and S is the salinity of the river plume at a given point. The iso-
plume density only depends on salinity and the salinity anomaly is
pleth (s¼ 0.05) is defined as the boundary between river plume
proportional to the density anomaly. Assuming that the plume
and ambient sea water. Geometrical characteristics of the fresh-
depth at the offshore edge is negligible with respect to the plume
water bulge and of the sediment distribution within the bulge (i.e.
depth at the coast and considering a linear relationship such that
the sediment bulge) are defined by means of the following vari-
∆ρ = β∆S, where ∆S is the salinity anomaly between river dis-
ables:
charge and ambient sea water and β = 0.79m3kg −1psu−1, the
∬ x ( So − S (x, y) ) dxdy ∬ y ( So − S (x, y) ) dxdy ∬ xC (x, y) dxdy ∬ yC (x, y) dxdy freshwater discharge of the alongshore coastal current for the
xf = ; yf = ; xc = ; yc =
∬ ( So − S (x, y) ) dxdy ∬ ( So − S (x, y) ) dxdy ∬ C (x, y) dxdy ∬ C (x, y) dxdy
Margules front can be computed as (Fong and Geyer, 2002):
⎛x ⎞ ⎛x ⎞
f
Rf = x 2 + y 2 ; Rc = x c2 + yc2 ; θ f = tan−1⎜⎜ ⎟⎟ ; θc = tan−1⎜⎜ c ⎟⎟ L
∬ v ∆SS0 dxdz= ∫0 ∆S
f f ⎝ yc ⎠ ρ0
⎝ yf ⎠ (2) Q fcc 0≡ v hdx≈ (g ′h0 )2
S0 2fgβS0 (4)
where S (x, y ) and C (x, y ) are surface salinity and suspended se-
diment concentration at a given point of the domain, xf and yf are where L is the alongshore coastal current width, h0 is the plume
126 H.Y. Yao et al. / Continental Shelf Research 116 (2016) 122–135

depth at the coast, and g′=g (∆ρ /ρ0 ) is the reduced gravitational h̅
acceleration without sediment. Eq. (4) suggests that the fresh- Tωs=
ωs (12)
water transport in the coastal current is a function of g′ and h0 and
it is thus dependent on density differences between river dis- where h̅ is the average freshwater plume depth. Here we use an
charge and ambient sea water. The lower the density difference, approximate value h̅ = 0.5 h0 .
We further introduce the non-dimensional number Y as the
the lower are g′and Q fcc0 . The above mentioned relationship has
ratio between these two timescales:
been used to estimate the alongshore freshwater transport in the
absence of any sediment input (Fong and Geyer, 2002). Tst ω
Y= =δ s
To account for density variations induced by the presence of Tωs fh0 (13)
sediments, we consider a modified plume density anomaly equal
3 + Fi2
to: where δ = 4π ( ) and ranges from 4π to 6π. The non-dimen-
2 + Fi2

⎛ρ ⎞ sional number Y represents the ability of the system to deliver


∆ρ′=∆ρ +Cr ⎜ r −1⎟ sediments to the alongshore coastal current. When Y > 1, sedi-
⎝ ρs ⎠ (5)
ments settle inside the recirculating bulge, and there is no sedi-
where ∆ρ′ is the density difference between ambient sea water ment flux in the coastal current. When Y < 1, sediments reach the
and river discharge in the presence of sediment, Cr is the sediment coastal current. Moreover, the smaller is the value of Y , the more
input concentration, ρr is the density of the input water (in the likely is for suspended sediment to be discharged into the along-
absence of sediments), and ρs is sediment density. The reduced shore coastal current.
gravitational acceleration in the presence of sediments, G′, can be A critical settling velocity value can be found by imposing the
written as: non-dimensional parameter Y equal to 1. To guarantee an along-
shore sediment transport ( Q scc ), it is thus necessary to have a se-

G′ =
∆ρ′
g=
∆ρ + Cr ( ρr
ρs
−1 )g diment settling velocity lower than the critical settling velocity
ρo ρo (6) value given by:

Laboratory experiments (Horner-Devine et al., 2006) suggest


fh0
ωsc =
δ (14)
that the distance of the bulge from the shoreline is proportional to
the inertial radius, L i , defined as: when this condition is fulfilled, an alongshore sediment transport,
v Q scc , forms in the coastal current, equal to:
L i= r
f (7)
Q scc ≡ ∬ vCdxdz (15)
where vr is the river flow velocity.
Another useful parameter is the inlet Kelvin number (Garvine, where C the suspended sediment concentration in the plume.
1987):

Bf
Ki= 4. Results
G′h0 (8)

where B is the river mouth width. When Ki is small, the rotation of 4.1. Effect of sediment dynamics on the alongshore freshwater
the plume flow is limited, but when Ki increases, the role of the current
earth centrifugal force becomes visible (Garvine, 1995).
According to Yankovsky and Chapman (1997), the bulge radius, The addition of fine sediments into the system leads to an in-
rs , and the azimuthal cyclostrophic velocity at the edge of bulge, v^s , creased density of the input water. An overview of the effect of
can be approximated by: sediments on the plume density field is given in Fig. 3 where the
density field is shown for four different experiments at the end of
2
G′h0 +v^s the fifth numerical day. Run A and B do not have any sediment
rs= −
^
fv s (9) input and their river discharges are characterized by 1 psu salinity
difference (28 and 29 psu respectively). On the contrary, salinity
and and sediment concentration in the river discharge are set to 28 psu
1 and 1.54 kg/m3 in run C and run D. Sediment settling velocity is
v^s=−( 2G′h0 + vi ) 2 (10) low for run C (10  5 mm/s), and equal to 0.1 mm/s for run D. Run B
To determine the farthest distance to which the sediment is and C are characterized by a very similar density field indicating
transported, two time scales are introduced here. The first time that differences due to salinity can be compensated by an increase
scale is the sediment transport timescale, Tst , defined as: or decrease in sediment input. In fact, the lower is the settling
velocity the more the system behavior approaches that of a system
2πrs 2π (3G′h0 + vi2 ) 2π (3+Fi2 ) having higher salinity values. This is due to the fact that low set-
Tst = = =
v^ (
f 2G′h0 + vi2 ) f (2+Fi2 ) tling velocities prevent deposition near the river mouth and allow
s (11)
the maintenance of higher densities in the entire flow field. Thus,
where Fi=vi/(G′h0 )1/2 is the inflow densimetric Froude number. The when settling velocity is low, sediments are uniformly spread and
sediment transport timescale determines the time taken by sedi- fill the entire bulge.
ments to move from the river mouth to the alongshore coastal The reduced alongshore freshwater transport ( Q fcc ) of run 3C
current traveling along the bulge edge. with respect to run 3A (Fig. 3) is due to the presence of sediments
The second significant timescale is the sediment settling in the river plume. To understand this process, the alongshore
timescale, Tωs , representing the average time taken by sediments to freshwater transport can be schematized as a simple two-layer
drop out from the river plume: Margules front:
H.Y. Yao et al. / Continental Shelf Research 116 (2016) 122–135 127

1023.5
A B C D
S r=28 psu S r=29 psu S r=28 psu S r=28 psu
340 3 3 3
Cr=0 kg/m Cr=0 kg/m Cr= 1.54 kg/m Cr=1.54 kg/m 3
-5 1023
s = 10 mm/s s =0.1 mm/s

1022.5
300

Desity(kg/m3 )
1022
Y(km)

260
1021.5
Q fcc=727m3 /s Q fcc=516m3 /s Q fcc=671m3 /s Q fcc=650m3 /s

1021
220

1020.5

180
1020
0 12 24 36 0 12 24 36 0 12 24 36 0 12 24 36
X(km) X(km) X(km) X(km)
Fig. 3. Surface density in the river plume for different salinities and sediment concentrations at t¼ 5 days: (A) Sr ¼28 psu, Cr ¼ 0 kg/m3; (B) Sr ¼ 29 psu, Cr ¼0 kg/m3; (C)
Sr ¼ 28 psu, Cr ¼1.54 kg/m3, ωs ¼ 0 mm/s and (D) Sr ¼ 28 psu, Cr ¼1.54 kg/m3, ωs ¼10  1 mm/s. For all results shown in Fig. 3, Q ¼ 10,000 m3/s, B ¼ 3 km, H ¼15 m. The
alongshore freshwater transport Q fcc is measured at y ¼ 255 km and decreases with increasing sediment input or increasing salinity.

ρ
∆ρ + C ( ρr −1) ∂h As a further simplification, we replace h0 with the river mouth
g ′h02 ⎛ ρr ⎞
Q fcc ≡ ∬ g
ρ0 f
s
∂x
dA≈Q fcc 0+ C ⎜ − 1⎟
2fβS0 ⎝ ρs ⎠
depth H. Hence, Eq. (18) becomes:
(16)
∆Q fcc =γCrm (19)
where Q fcc is the freshwater discharge computed in the presence
suspended sediment and Q fcc0 the freshwater discharge without with
ρ
sediments. On the RHS of Eq. 16, ( ρr −1) is always negative ( ρr < ρs )
and it is thus representative of a decrease in the freshwater
s

γ =α
( ) ρ
g ′ 1− ρr H 2
s

transport dictated by the increased flow density. Hence, the dif- 2fβ S 0 (20)
ference in alongshore freshwater transport with and without the
where α , γ ,, and m are constant coefficients. A regression based on
addition of sediments is:
model results gives γ=35. 97 and m ¼1.66 with R2 ¼ 0.995
g ′h02 ⎛ ρ ⎞ (p o0.05). Therefore, increasing sediment input concentration
Q fcc 0−Q fcc =∆Q fcc ≈ C⎜1 − r⎟ traps more freshwater in the circulating bulge and reduces the
2fβ S 0 ⎝ ρs ⎠ (17)
alongshore coastal current (Fig. 4B).
For a constant salinity value and Coriolis parameter, the re-
duction in freshwater transport is a function of the plume depth at 4.2. Effect of sediment dynamics on the freshwater bulge
the coast h0 and of the suspended sediment concentration C .
Assuming C in Eq. (17) as proportional to the river suspended se- A series of numerical experiments were conducted to in-
diment concentration Cr , the reduction in alongshore freshwater vestigate the effect of an increase in sediment input on the dy-
transport becomes: namics and structure of the recirculating bulge (Fig. 5). Colored
∆Q fcc ∝Cr h02 (18) points represent the decline of θf (Fig. 5A) and Rf (Fig. 5C) as a
function of sediment input concentration. The center of the
Based on the models results measured at t¼ 5 days, Fig. 4A freshwater bulge moves closer to the shoreline and θf slightly
shows the relationship between Cr h02 and the reduced freshwater decreases when the sediment concentration increases (Fig. 5A).
transport (geometric interpolation R2 ¼0.99, p o0.05). The in- The increased plume rotation also causes a slightly decrease in Rf ,
crease of h0 is proportional to Cr , indicating that sediment con- with more freshwater trapped in the recirculating bulge and a
centration is a controlling parameter and regulates variations of reduction of the alongshore freshwater current. These results are
freshwater transport in the coastal current (Fig. 4B). consistent with Figs. 3 and 4, where the freshwater discharge in
128 H.Y. Yao et al. / Continental Shelf Research 116 (2016) 122–135

50 A 40 B

35
40
30

25
ΔQfcc ( m3/s )

ΔQfcc ( m3/s )
30
20
20 15

10
10
5

0 0
0 500 1000 0 0.5 1 1.5
Crh20 ( kg/m3 ) Cr ( kg/m3 )
Fig. 4. Relationships between (A) Reduction in freshwater transport ∆Q fcc and Cr h02, (B) Reduction in freshwater transport ∆Q fcc as a function of the sediment input
concentration. Blue points are results from the numerical model for t ¼ 5 days, Q ¼ 10,000 m3/s, Sr ¼ 28 psu, B ¼3 km, H ¼ 15 m, Cr varying from 0.1 to 1 kg/m3. (For
interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

35 50
30 40
θf ( ° )

θf ( °)

25 30
20 A B
20
0 0.5 1 0 0.4 0.8
3 v i (m/s)
C r (kg/m )

1.6 10
4 10 4
2
Rf ( m )

Rf ( m )

1.3 1.5

C D
1 1
0 0.5 1 0 0.4 0.8
C r (kg/m 3 ) v i (m/s)
Fig. 5. (A) Distance of the bulge center from the river mouth (Rf ) as a function of sediment input concentration (Cr ). (C) Angle with respect to the shoreline θf as a function of
sediment input concentration. Numerical results refer to Q ¼ 10,000 m3/s, Sr ¼28 psu,t = 5 days, B ¼ 3 km, H ¼15 m and Cr varying from 0.1 kg/m3 to 1 kg/m3. (B) Distance of
the bulge center from the river mouth (Rf ) as a function of river flow velocity. (D) Angle with respect the shoreline (θf ) as a function of river flow velocity. Numerical results
refer to t ¼ 5days, river discharge ranging from 2500 m3/s to 30,000 m3/s, Cr ¼ 1 kg/m3, Sr ¼28 psu, B ¼ 3 km, H ¼15 m. Gray lines represent linear interpolations.

the alongshore current declines with increasing sediment input of sediment input concentration (Fig. 5B and D). Increasing river
concentration. flow velocity from 0.056 m/s to 0.67 m/s causes θf and Rf to in-
The influence of river flow velocity on the structure of the re- crease by 70% and 38% respectively (Fig. 5B and D), while varia-
circulating bulge is more significant with respect to the influence tions in sediment input concentration cause a decrease in θf and Rf
H.Y. Yao et al. / Continental Shelf Research 116 (2016) 122–135 129

315 C
× 104
1.9 A
R f (m) 1.7

1.5 ωs=1mm/s

1.3
10-3 10-2 10-1 100

Y(km)
290
ωs (mm/s) -4
ωs=10 mm/s
50 B ωs=0.1mm/s
40
θf (°)

30

20
10-3 10-2 10-1 100
265
12 24 36
X(km)
ωs (mm/s)

Fig. 6. (A) Distance Rf of the bulge center from the river mouth and (B) angle θf with the shoreline as a function of sediment settling velocity, for 3 different river discharges
measured at t ¼5 days. Blue line Q ¼ 5000 m3/s; pink line Q ¼ 10,000 m3/s; green line Q ¼ 20,000 m3/s. (C) Surface salinity boundary at t¼ 5 days, Q ¼ 10,000 m3/s, red line
ωs ¼ 10  4 mm/s, green line ωs ¼0.1 mm/s, blue line ωs ¼1 mm/s. All points have, Cr ¼ 1 kg/m3, Sr ¼ 28 psu, B ¼3 km, H ¼ 15 m and ωs varies from 10  3 mm/s to 1 mm/s. (For
interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

smaller than 10%. This is in agreement with the experimental re- velocities are higher than 10  2 mm/s (Fig. 6A and B). In fact, once
sults of Horner-Devine et al. (2006), indicating that the distance of the settling velocity is above10  2 mm/s, the possible increase in
the freshwater bulge from the shoreline is mainly proportional to bulge rotation (due to an increase in fluid density) is canceled out
the inertial radius (Eq. (7)). by the fact that sediments rapidly drop out from the river plume
To further explore the effect of sediments on bulge hydro- depositing near the river mouth, without resulting in an increase
dynamics, simulations with multiple settling velocities were con- in fluid density. The freshwater bulge is thus less flattened against
ducted (Fig. 6). The planar structure of the freshwater bulge is the alongshore side of the domain (Fig. 6C) and the freshwater
slightly affected by the sediment settling velocity when settling bulge angle θf slightly increases.

50 50
A B
θc ( ° )

θc ( ° )

40 40

30 30

0 0.5 1 0 0.5
C (kg/m 3 ) v i (m/s)

10 4 10 4
2 2
C D
R c (m)

R c (m)

1.5 1.5

1 1

0 0.5 1 0 0.5
3 v i (m/s)
C (kg/m )
Fig. 7. (A) Angle θc as a function of input sediment Concentration for two different settling velocity at t¼ 5 days. Q ¼ 10,000 m3/s, Sr ¼28 psu, B ¼ 3 km, H ¼ 15 m and Cr vary
from 0.1 kg/m3 to 1 kg/m3. (C) Distance Rc of the sediment distributional center from the river mouth as a function of different sediment input concentrations and for two
different settling velocities. (B) Angle θc as a function of river flow velocity. (D) Distance Rc of the sediment distributional center from the river mouth. Points are taken at
t ¼ 5 days, with river discharge ranging from 2500 m3/s to 30,000 m3/s ,Cr ¼1 kg/m3, Sr ¼28 psu, B ¼3 km, H ¼ 15 m. Blue and red points refer to settling velocities equal to
10  2 and 10  4 mm/s respectively. Gray lines represent linear interpolations. (For interpretation of the references to color in this figure legend, the reader is referred to the
web version of this article.)
130 H.Y. Yao et al. / Continental Shelf Research 116 (2016) 122–135

4.3. Sediment distribution within the bulge Similarly to Rc and Rf , the difference between θf and θc decreases
with increasing river discharge. Moreover, with increasing settling
Sediment distribution within the freshwater bulge is dictated velocities the angle θc approaches 90°. This behavior can be ex-
by the hydrodynamic characteristics of the system, such as flow plained considering the sediment's residence time within the
velocity, density anomaly, plume thickness, as well as by sediment bulge. Residence time increases with decreasing settling velocity.
properties, such as sediment settling velocity. Fig. 7A and C show The lower the settling velocity, the longer the sediment stays in
the relationship between sediment bulge geometrical character- the plume and the more likely it will be transported far from the
istics (center of mass, Rc , angle between the shoreline and the line river mouth. On the other hand, when the settling velocity is high,
connecting the river mouth and the center of mass, θc ) and sedi- there is no time for the Coriolis force to affect sediment transport
ment input concentration. Fig. 7B and D show the relationship and for this reason the angle of the sediment bulge approaches
between sediment bulge geometrical characteristics and river 90°when Rc decreases.
mouth velocity.
Increasing sediment input concentration causes Rc and θc to 4.4. Sediment transport in the coastal current
slightly decrease. The same decrease has been observed for the
freshwater bulge parameters θf and Rf (Fig. 5A and C). This suggests Herein we investigate the effect of sediment input, settling
that changes in the reduced gravitational acceleration affect the velocity, and river discharge on the sediment transport in the
distribution of suspended sediment as well, causing the distance alongshore current.
of the center of mass of suspended sediment from the river mouth
and its angle with respect to the shoreline to decrease (Fig. 7A and 4.4.1. Effect of sediment concentration and settling velocity on the
C). River flow velocity (Fig. 7B and D) strongly affects sediment alongshore sediment transport
distribution, with high river flow velocities transporting and de- Sediment transport ( Q ssc ) and freshwater transport ( Q fcc )
positing suspended sediments farther from the river mouth. In this measured at y¼ 255 km and at the end of the fifth numerical day
case flow inertia prevails over rotational forces, so that both Rc and are presented in Fig. 9A as a function of river discharge (see Fig. 3
θc increase with increasing river velocity. These results confirm for the location of the cross section). Discharge values have been
what was found analytically and from in situ and satellite ob- varied between 2500 m3/s to 30,000 m3/s, while maintaining a
servations by Falcini et al. (2012b) for the Mississippi River. constant input of sediments and a settling velocity equal to
The structure of the suspended sediment bulge depends on the 10  4 mm/s. The small settling velocity value was chosen to facil-
sediment settling velocity, with the sediment bulge differing from itate the alongshore sediment transport. Q ssc increases with in-
the freshwater bulge once the settling velocity is high (Figs. 2 and creasing river discharge Q . This is due to the fact that the fresh-
3C and D). Fig. 8A shows the effect of sediment settling velocity on water discharge in the coastal current Q fcc increases. Moreover,
the distance of the sediment center of mass from the river mouth, Q ssc and Q fcc growth rates are higher when the river discharge is
Rc , for different discharge values (5000, 10,000 and 20,000 m3/s). less than 20,000 m3/s.
The distance Rc remains constant for low settling velocities and Fig. 9B shows numerical results for a series of simulation
thus displays a trend similar to the freshwater bulge distance Rf runs with increasing sediment input concentration (sediment
(see Fig.6). Differences between the distance of the freshwater input ranging from 0.1 to 4 kg/m3) and constant discharge
bulge from the river mouth Rf and the sediment bulge distance Rc ( 10,000 m3/s). Increasing sediment concentration augments se-
are present due to early sediment deposition. The sudden slope diment availability in the whole model domain and it also en-
change in Fig. 8A indicates rapid sediment deposition within the hances sediment transport within the coastal current. As discussed
bulge for ωs >10  2 mm/s. The angle of the sediment bulge center θc in the previous sections, increasing the density of the plume water
remains constant for low settling velocity, and maintains thus a decreases the freshwater transport within the alongshore current,
similar trend with respect to the angle of the freshwater bulge θf . and thus reduces the corresponding sediment transport. Therefore
a critical input concentration exists that maximizes suspended
sediment transport in the coastal current. Concentrations higher
than the critical value are unfavorable to alongshore sediment
transport (Fig. 9B, the blue line reaches a maximum and then
decreases). Fig. 9C shows numerical model results for different
settling velocities, ranging from 10  3 mm/s to 5 mm/s. The
alongshore sediment transport Q ssc remains constant when the
settling velocity is less than 10  2 mm/s. For settling velocities
higher than 10  2 mm/s, the sediment transport Q ssc decreases.
Sediment transport becomes negligible for settling velocities
above 0.1 mm/s, indicating that all the sediment is deposited in
the recirculating bulge before entering the alongshore current.
Fig. 10 represents the ratio of sediment discharge in the coastal
current ( Q ssc ) to sediment discharge at the river mouth ( Q ssr ) as a
function of the non-dimensional number Y. The sediment dis-
charge ratio is stable for Yo0.01 and then rapidly decreases. For
Y41 there is no sediment flux in the costal current and all sedi-
ments discharged from the river are deposited in the bulge area.
This is because the travel time along the bulge is longer than the
Fig. 8. (A) Distance Rc of the sediment distributional center from the river mouth time necessary to sediments to drop out from the bulge itself.
and (B) angle θc as a function of sediment settling velocity and for different river
discharges. Colors green, pink and blue represent river discharge of 20,000 m3/s, 4.4.2. Effect of discharge intensity on the alongshore sediment
10,000 m3/s and 5000 m3/s respectively. For all results plotted in this figure, Cr
¼ 1 kg/m3, Sr ¼28 psu, B ¼ 3 km, H ¼ 15 m and ωs vary from 10  3 mm/s to 1 mm/s,
transport
t¼ 5 days. (For interpretation of the references to color in this figure legend, the In this paragraph we investigate the effect discharge intensity
reader is referred to the web version of this article.) on the alongshore sediment transport. Fig. 11 represents the ratio
H.Y. Yao et al. / Continental Shelf Research 116 (2016) 122–135 131

20 A 2000

Qfcc (m3/s)
Qssc (t/s)
10 1000

0 0
0 0.5 1 1.5 2 2.5 3
Q (m3/s) x 10
4

12 B 1000

Qfcc (m3/s)
Qssc (t/s)
9
6 500
3
0 0
0 0.5 1 1.5 2 2.5 3 3.5 4
C (kg/m3)
6 700

Qfcc (m /s)
690
Qssc (t/s)

3
4 680
670
2 660
650
0 640
−5 −4 −3 −2 −1 0
10 10 10 10 10 10
(mm/s)
s

Fig. 9. Variations in freshwater transport (blue line) and suspended sediment transport (green line) in the alongshore current as a function of (A) river discharge,
(B) suspended sediment concentration, (C) sediment settling velocity. (A) Cr ¼1 kg/m3, ωs ¼ 10  4 mm/s, Q ranges from 2500 m3/s to 30,000 m3/s. (B) Q ¼ 10,000 m3/s, ωs
¼ 10  4 mm/s, Cr ranges from 0.1 kg/m3 to 4 kg/m3 (C) Cr ¼ 1 kg/m3, Q ¼ 10,000 m3/s, ωs ranges from 10  4 mm/s to 5 mm/s. All results are measured at the end of day 5 at
y¼ 255 km with Sr ¼ 28 psu, B ¼ 3 km, H¼15 m. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

of alongshore sediment transport to the sediment flux at the notice, the alongshore sediment transport increases with in-
mouth as a function of input river discharge. Panel A shows results creasing river discharge for Fi <1. This can be explained by the
using a lower settling velocity with respect to panel B. The corresponding decrease of the non-dimensional Y parameter.
alongshore sediment transport was calculated at the end of the However, once the flow becomes supercritical ( Fi >1), the along-
5th numerical day. Gray areas in both panels indicate a subcritical shore sediment transport drastically decreases. This is due to the
river flow ( Fi <1). Filled circles indicate simulations characterized fact that with a supercritical flow the bulge becomes unstable after
by a stable bulge, open circles indicate unstable runs (bulge peri- 5–6 rotations (see also Fig. 2). When the bulge is unstable, the flow
odically attaching and detaching from the wall). As it is possible to periodically detaches from and then reattaches to the wall. Bulge

0.6

0.5

0.4
Qssc /Qsscr

0.3

0.2

0.1

0
10-4 10-3 10-2 10-1 100 101 102
Y

Fig. 10. Ratio between sediment discharge in the coastal current measured at y ¼ 255 km and sediment discharge at the river mouth, as a function of the ratio Y between the
sediment transport time scale and the sediment settling time scale. Circles: Q ¼ 10,000 m3/s, ωs ¼ 10  4 mm/s and Cr ranges from 0.1 kg/m3 to 1 kg/m3; triangles: Q ¼
10,000 m3/s, ωs ¼10  2 mm/s, and Cr ranges from 0.1 kg/m3 to 1 kg/m3; asterisks: Q ¼ 20,000 m3/s, Cr ¼ 1 kg/m3and ωs ranges from 10  3 mm/s to 5 mm/s; plus signs: Q ¼
10,000 m3/s, Cr ¼1 kg/m3and ωs ranges from 10  3 mm/s to 5 mm/s; crosses: Q ¼ 5000 m3/s, Cr ¼ 1 kg/m3and ωs ranges from 10  3 mm/s to 5 mm/s. All results have
Sr ¼ 28 psu, B ¼3 km, H¼15 m, t¼ 5 days.
132 H.Y. Yao et al. / Continental Shelf Research 116 (2016) 122–135

0.25 Fi < 1
given a certain sediment input concentration and settling velocity,
Fi
a critical discharge value exists which maximize the alongshore
Qssc / Qsscr

0.20 sediment transport. This critical discharge value corresponds to


the limit between subcritical and supercritical flow.
0.15 Following the definition of the non-dimensional Y number, to
guarantee an alongshore sediment transport ( Q scc ), it is necessary
0.10 to have sediment settling velocity values lower than a critical
settling velocity (Eq. (14)). Fig. 12A shows critical settling velocity
0.05 values, ωsc , as a function of the input river discharge. In the sub-
A critical case (Fi o1) for a given sediment input concentration, the
0 higher is the river discharge the higher is the value of the critical
0 1 2 3 4 5 6
settling velocity. In the figure multiple values for individual dis-
4 3
Qr (x 10 m /s) charges are present because the different parameters used in the
simulations affect the bulge structure. When the flow is super-
0.04 critical (Fi 41) the bulge becomes unstable, with a sharp reduc-
Fi < 1 Fi
tion in alongshore freshwater transport (Fig. 2D). This translates in
Qssc / Qsscr

0.03 a sudden decrease in critical settling velocity (Fig. 12A), with most
sediment fractions trapped in the detached bulge.
These results suggest that, in a subcritical flow, low river dis-
0.02
charge values promote sediment sorting in the inner shelf since only
the finest sediment fractions are transported in the alongshore
0.01 coastal current. To better explain this concept, Fig. 12B and C show
the instantaneous (black lines) sediment transport in the alongshore
0
B coastal current, and its cumulative (red lines) values for two simu-
0 1 2 3 4 5 6 lations run for low and high river discharge values and for two dif-
4 3 ferent settling velocities. For the high river discharge test, the total
Qr (x 10 m /s) amount of sediments (red line) transported in the alongshore coastal
Fig. 11. Ratio between sediment input concentration and sediment transport in the current becomes stable around the 5th day, while it takes 12 days for
alongshore coastal current as a function of river discharge (A) Cr ¼ 1 kg/m3, the low river discharge case to reach stability. Thus, despite of the
ωs ¼ 0.05 mm/s and (B) Cr ¼ 1 kg/m3 ,ωs ¼0.1 mm/s. Gray areas are zones of sub- increasing bulge size, sediments reach the alongshore coastal current
critical flow.
more easily in case of high river discharge. This can be explained by
considering that, for given concentration and settling velocity values,
instability results in a pulsing transport in the alongshore coastal the Y parameter decreases with increasing discharge (Fig. 12A). It is
current. This pulsing period is around 5 bulge rotations, and then possible to notice that for both the high and low river discharge
eventually causes a decrease in the corresponding alongshore conditions, the finest sediment fractions are more easily transported
transport. Our results are in agreement with laboratory experi- in the alongshore coastal current (in both plots, the continuous red
ments and numerical tests reported in Horner-Devine et al. (2005). line is always higher than the dashed line). However, such differences
Similar detachment processes are also found in the Loop Current, in the transport of fine and coarse sediments are much more pro-
Gulf of Mexico (Nof and Pichevin, 2001), and Agulhas Current, nounced in case of low river discharge (see differences between
southwest Indian Ocean (Pichevin et al., 1999). Due to bulge in- continuous and dashed lines in Fig. 12B and C). This differential
stability, and intermittency in the alongshore transport, for Fi >1 transport rate between different sediment fractions indicate that low
the Y parameter cannot be considered representative for the and long lasting flow discharge promote sediment sorting in the
alongshore sediment transport anymore. These results show that, continental shelf.

Fig. 12. (A) Critical settling velocity computed from the numerical simulations as a function of river discharge. River discharge has been varied from 5000 to 50,000 m3/s.
Input concentration is equal to 1 kg/m3, sediment settling velocity were varied from 10  4 to 1 mm/s. Multiple critical settling velocity values correspond to one river
discharge value because different settling velocities change the structure of the freshwater bulge. (B) Sediment flux in the alongshore coastal current (black lines) and
cumulative fluxes (red lines) for Cr ¼ 1 kg/m3, Q ¼ 30,000 m3/s, and for two different settling velocities: ωs ¼ 10  2 mm/s and ωs ¼ 10  4 mm/s. (C) Sediment flux in the
alongshore coastal current (black lines) and cumulative fluxes (red lines) for Cr ¼1 kg/m3, Q ¼ 6000 m3/s, and for two different settling velocities: ωs ¼ 10  2 mm/s and ωs
¼ 10  4 mm/s. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
H.Y. Yao et al. / Continental Shelf Research 116 (2016) 122–135 133

Q (m3/s) ×10
4 (Liu et al., 2004) and the Amazon River in Brazil (Allison et al.,
3 1.5 1 0.75 0.6 0.5 2000; Anthony et al., 2010). A clinoform deposit has been also
0.7
ωs= 0. 0001mm/ s reported in a coastal strip extending 800 km from the Yangtze
ωs= 0. 01mm/ s Estuary in China (Liu et al., 2006; Milliman et al., 1985b). Although
TotalQssc /TotalQ sscr

0.6 riverine sediments can be dispersed along the coastline by waves


and currents far beyond the freshwater plume, the transport of
0.5 very fine sediment fractions in trapped freshwater plumes can be
significant, thus affecting the subaqueous morphology of the in-
ner-shelf. Herein we mainly focus on the effect sediment char-
0.4
acteristics, river discharge, flood durations and flood intensity on
sediment dispersal.
0.3
0 2 4 6 8 10 12
Duration (days) 5.1. Effect of suspended sediment concentration, settling velocity, and
flocculation on alongshore sediment transport
Fig. 13. Ratio of alongshore sediment transport to sediment flux at the mouth for
different flood durations and settling velocities. Duration ranges from 2 days to 10
days to guarantee the same input volume of freshwater and sediment. Red points:
The suspended sediment transport in the alongshore coastal
ωs ¼ 0.05 mm/s, Q ranges from 6000 m3/s to 30,000 m3/s; blue points: ωs ¼0.1 mm/ current depends on freshwater discharge and related suspended
s, Q ranges from 6000 m3/s to 30,000 m3/s. (For interpretation of the references to sediment concentration. Increasing the suspended sediment con-
color in this figure legend, the reader is referred to the web version of this article.) centration discharged by the river has a twofold effect on the se-
diment flux along the shoreline. On one hand it triggers a decline
in freshwater transport in the alongshore coastal current, possibly
4.4.3. Effect of discharge duration on the alongshore sediment
decreasing the delivery of sediments in the coastal current itself
transport
(Eq. (19), Fig. 4). On the other hand, increasing sediment con-
Natural rivers are characterized by floods of different duration.
centration obviously increases sediments availability within the
Herein, the effect of flood duration on the alongshore sediment
system. This twofold effect results in a critical value of suspended
transport is investigated. For the set of simulations presented in
sediment concentration that maximizes sediments delivery to
this paragraph, an identical volume of water and sediments is
the coastal current (Fig. 9B). It is important to note that only very
delivered to the ocean. Flood duration ranges from 2 to 10 days,
high sediment concentrations reduce sediment delivery (above
and discharge values range from 6000 m3/s to 30,000 m3/s.
2.5 kg/m3 in Fig. 9B). These concentrations likely trigger floccula-
The input sediment concentration has been set constant and
equal to 1 kg/m3. Two different settling velocity values, equal to tion, reducing the total sediment volume in suspension and
0.0001 mm/s and 0.01 mm/s have been adopted. The total amount changing the bulge dynamics. Flocculation was not accounted for
of sediment delivered to the alongshore coastal current was in these simulations.
computed at the end of the 20th numerical day for each simula- For a given a sediment input concentration, information about
tion (Fig. 13). Therefore, after flood occurrence, the numerical si- the transport of a certain sediment fraction within the coastal
mulation continued with no freshwater or sediment input. current can be obtained by means of the non-dimensional para-
In the first part of the plot the sediment transport in the meter Y (Eq. (13)). For Y4 1, no sediment reaches the alongshore
alongshore coastal current increases with flood duration. This is coastal current; for Y o1, an alongshore sediment flux is present.
due to the temporal evolution of both freshwater bulge and The lower is Y , the higher is the likelihood for sediments to be
alongshore coastal current. In fact, during the first 3–4 rotation transported into the alongshore coastal current. By setting the
periods the alongshore transport increases, and it remain constant parameter Y to one it is possible to calculate the maximum settling
thereafter (Horner-Devine et al., 2006). Therefore up to 4 days, the velocity value above which the sediment is not allowed to enter
establishment and development of the alongshore coastal current the alongshore coastal current. As an example, based on the si-
lead to an increasing alongshore sediment transport. After 4 days, mulations conducted in this paper for a discharge of 10,000 m3/s,
the alongshore coastal current in fully developed. To maintain a maximum settling velocities yielding an alongshore sediment
constant input of water and sediments in the system, we de- transport range from 0.07 mm/s to 0.14 mm/s. Coarser sediment
creased river discharge values with increasing discharge duration. fractions will deposit within the bulge before reaching the coastal
This decrease in river discharge leads to a decrease in the along- current.
shore sediment transport as reported in the second part of the As mentioned before, when sediment concentrations are very
plot. Therefore, given a certain input volume of sediment and high, flocculation processes can be important. However, even
water into the system, a critical discharge duration exists which when flocculation and turbidity maxima are present, part of the
maximizes the alongshore sediment transport. The existence of sediment load is transported far from the river mouth. Floccula-
such critical discharge duration is due to the temporal develop- tion triggers an increase in settling velocity, resulting in a larger
ment and establishment of the alongshore coastal current which freshwater bulge (Fig. 6c). However, the corresponding sediment
requires around 4 complete bulge rotations to be fully established. bulge will be smaller and closer to the river mouth because sedi-
ments drop out of the plume at a faster rate (Fig. 8). Increasing the
size of the freshwater bulge reduces the amount of freshwater able
5. Discussion and conclusions to reach the coastal current, so that the freshwater discharge in the
coastal current decreases when we increase settling velocity
Coarse sediment fractions and aggregated flocs are generally (Fig. 9c). The sediment load in the coastal current is also reduced,
rapidly deposited near the river mouth. Only disaggregated, clay- both because more sediment is deposited in the bulge and because
sized particles can be transported long enough to enter the of the flow reduction (Fig. 9c). In conclusion, the formation and
alongshore coastal current (Geyer et al., 2004a). Elongated coastal settling of large flocs near the river mouth affects the amount of
mud wedges formed by alongshore sediment transport have been water reaching the coastal current, and further reduce the trans-
observed near some large rivers, such as the Yellow River in China port of the un-flocculated fraction away from the river mouth.
134 H.Y. Yao et al. / Continental Shelf Research 116 (2016) 122–135

5.2. Effect of sediments on the bulge structure subcritical zone, the increase in alongshore sediment transport
with increasing discharge can be explained by considering that the
The addition of sediments to a surface-trapped riverine plume Y parameter decreases with river discharge (Eq. (13)). Decrease in
system causes a reduction in density anomaly, which in turn af- the Y parameter suggests that, as long as the flow is subcritical, an
fects the freshwater transport in the alongshore coastal current. increase in flow velocity (and thus an increase in the velocity of
The Kelvin number, which is the ratio of outflow width and in- sediments traveling along the edge of the bulge), compensates the
ternal Rossby number, can be used as representative for the earth increase in bulge size. Once the flow becomes supercritical ( Fi >1),
rotational effects (Odonnell, 1990; Wiseman and Garvine, 1995). the bulge becomes unstable and the location where the outer
The reduced gravitational acceleration ( G′) decreases with in- bulge impinges on the wall moves upstream cutting off the coastal
creasing sediment input thus increasing the Kelvin number (Eq. current (Fig. 2D, see also Horner-Devine et al., 2005). Thus, once
(8)). As a result, an increase in sediment input and consequent the flow becomes supercritical the alongshore transport becomes
increase in fluid density enhances the rotational effect of the intermittent and consequently decreases (Fig. 11). This effect is
Coriolis force with respect to river inertial forces. more pronounced for high settling velocities (Fig. 11B).
The response of the system to these density variations strongly Model results indicate that for a certain discharge there is a
depends on sediment settling velocity and river discharge values. critical settling velocity above which no sediments can be trans-
When the settling velocity is low (ωs o10  2 mm/s), the addition of ported in the alongshore coastal current. In a subcritical flow,
sediment has an effect similar to an increase in the salinity of the critical settling velocity values increase with river discharge
riverine plume. High sediment concentrations tend to move the (Fig. 12A). Therefore, long lasting and low magnitude flood events
center of the freshwater bulge towards the river mouth and to- maximize sediment sorting in the inner shelf, because large vo-
wards the shoreline. However, the effect of sediment concentra- lumes of coarse sediments are deposited near the river mouth
tion is minimal when compared to changes induced by variations while only the finest sediment fractions move forward in the
in river velocity. In fact, the position of the freshwater bulge is also alongshore current (Fig. 12B and C). An example of river char-
affected by different discharge values. This effect can be described acterized by long lasting floods is the Yangtze River, for which the
considering that the inertial radius ( L i ), which controls the dis- seasonal flood usually lasts over 3–4 months. During this period,
tance of the bulge center from the shoreline, is proportional to the 75% of the total annual sediment load is delivered to the ocean
river flow velocity (Horner-Devine et al., 2006). Our results are (McKee et al., 1983). Sediment sorting is evident in this system:
also in agreement with the numerical simulations of Marsalaix coarse grain and flocculated suspended sediment are trapped in
et al. (1998), showing that the momentum advection is responsible both the sub-aerial and sub-aqueous parts of the Yangtze delta
for the large offshore extension of the plume. (Chen et al., 2000), while the finest sediment fractions accumulate
With increasing settling velocity, most of the sediment drops along the coast (Liu et al., 2006; Liu et al., 2007; Milliman et al.,
out of the freshwater plume, does not reach the edge of the re- 1985a).
circulating bulge, and does not reduce the density anomaly. In this In case of episodic, high-magnitude floods with duration of the
case, the presence of sediments can no longer be compared to a order of days, high river discharge leads to the formation of a large
salinity increase. For a high concentration of sediments with high bulge while increasing the critical settling velocity. Therefore,
settling velocity, a large bulge radius is needed so that the cen- there is an enhancement of the seaward dispersal of suspended
trifugal force balances the Coriolis force (Eq. (9)). Hence, a reduc- sediment and transport of both fine and coarse sediment fractions
tion in settling velocity decreases the offshore growth of the re- far from the river mouth. This result is in good agreement with
circulating bulge (Fig. 6C). field data from the Eel River (Sommerfield and Nittrouer, 1999).
The distribution of sediments in the bulge largely depends on The 90% of the annual Eel River discharge occurs in two months
settling velocity. For the lightest sediment fractions, since θc re- due to intensive rainstorms (Brown and Ritter, 1971).This high
mains almost constant for different settling velocities and river magnitude events result in multiple sediment fractions dispersed
discharges, the inertial radius L i can be considered as a first proxy along the entire continental shelf and in adjacent regions of the
for Rc . Since the Coriolis parameter f is constant, river flow ve- Northern California continental Margin (Sommerfield and Nit-
locity is the only varying parameter in Eq. (7). As a consequence, Rc trouer, 1999).
increases with increasing discharge, which is in agreement with Finally, given the same input volume of water and sediments, a
results shown in Fig. 7D. For coarser sediment fractions, Rc largely critical flood duration exists that maximizes the alongshore
depends on settling velocity rather than riverine discharge. Hence, freshwater transport (Fig. 13). The critical flood duration is around
when the settling velocity is low ( ωs o10  2 mm/s), sediments fill 4 days and is connected to the temporal development of both
the entire recirculating bulge, and the location and size of the freshwater bulge and coastal current, which need to be fully es-
sediment bulge is largely affected by the riverine flow. When the tablished in order to maximize alongshore transport.
settling velocity is high ( ωs 410  2 mm/s), sediments are unable to Many processes, which have been neglected in the present
fill the entire freshwater bulge, and both Rc and θc linearly de- work, could affect the system behavior. Among others, wind
crease with settling velocity. In this case the location of the sedi- waves, tides, currents, and gradients in water temperature could
ment bulge is similar for different river discharges. have a non-negligible effect. For instance, tidal currents have been
found to stabilize the offshore growth of the river plume bulge and
5.3. Effect of flood duration and intensity on alongshore sediment to regulate the freshwater discharge in the alongshore coastal
transport current (Isobe, 2005). Bottom topography can affect the plume
hydrodynamics as well (e.g. Marsaleix et al., 1998). Moreover,
The fate of riverine sediments discharged into the shelf is winds alter river plumes: a downwelling wind tends to increase
highly affected by the intensity and duration of flood events. In the downcoast transport while an upwelling wind can move the
this manuscript, we explored the influence of flood duration and freshwater plume offshore (Hetland and Hsu, 2013). Moreover, in
magnitude on the alongshore sediment transport. Given the same this manuscript we did not account for the morphological evolu-
discharge duration, and for increasing river discharges, the tion of the seafloor and for sediment exchange between river
alongshore sediment transport increases up to the point where the plume and sea bottom caused by sediment re-suspension. In a
supercritical flow condition is reached, at that point the along- natural coastal environment, suspended sediments tend to settle
shore sediment transport start decreasing (Fig. 11). In the in the near field area within the freshwater bulge. The sediment is
H.Y. Yao et al. / Continental Shelf Research 116 (2016) 122–135 135

then re-suspended by energetic waves and currents and delivered experiments simulating a coastal river inflow. J. Fluid Mech. 555, 203–232.
further offshore (Traykovski et al., 2000; Wright et al., 2001). This Isobe, A., 2005. Ballooning of river-plume bulge and its stabilization by tidal cur-
rents. J. Phys. Oceanogr. 35, 2337–2351.
process repeats multiple times until the sediment is carried Krone, R.B., 1962. Flume Studies of the Transport of Sediment in Estuarial Shoaling
to a water depth large enough to prevent mobilization (Falcini Processes.
Leonardi, N., Canestrelli, A., Sun, T., Fagherazzi, S., 2013. Effect of tides on mouth bar
et al., 2012a). morphology and hydrodynamics. J. Geophys. Res.: Oceans 118 (9), 4169–4183.
All simulations presented herein were carried out with a Kelvin Leonardi, N., Kolker, A.S., Fagherazzi, S., 2015. Interplay between river discharge and
inlet number between 0.3 and 0.48, corresponding to intermediate tides in a delta distributary. Adv. Water Resour. 80, 69–78.
Lentz, S.J., Helfrich, K.R., 2002. Buoyant gravity currents along a sloping bottom in a
conditions between small-scale discharges (strong non-linear flow rotating fluid. J. Fluid Mech. 464, 251–278.
and limited Coriolis effect) and large-scale discharges (linear dy- Lesser, G.R., Roelvink, J.A., van Kester, J.A.T.M., Stelling, G.S., 2004. Development and
namics and across-shore geostrophic balance, see Garvine 1995). validation of a three-dimensional morphological model. Coast. Eng. 51,
883–915.
In future research it would be interesting to explore how large Li, J.F., Zhang, C., 1998. Sediment resuspension and implications for turbidity
variations in Kelvin inlet number (and therefore inlet geometry) maximum in the Changjiang Estuary. Mar. Geol. 148, 117–124.
Liu, J.P., Li, A.C., Xu, K.H., Veiozzi, D.M., Yang, Z.S., Milliman, J.D., DeMaster, D., 2006.
might affect the dynamics of suspended sediments in river Sedimentary features of the Yangtze River-derived along-shelf clinoform de-
plumes. posit in the East China Sea. Cont. Shelf Res. 26, 2141–2156.
Liu, J.P., Milliman, J.D., Gao, S., Cheng, P., 2004. Holocene development of the Yellow
River subaqueous delta, North Yellow Sea. Mar. Geol. 209, 45–67.
Liu, J.P., Xu, K.H., Li, A.C., Milliman, J.D., Velozzi, D.M., Xiao, S.B., Yang, Z.S., 2007. Flux
Acknowledgments and fate of Yangtze river sediment delivered to the East China Sea. Geomor-
phology 85, 208–224.
Lohrenz, S.E., Fahnenstiel, G.L., Redalje, D.G., Lang, G.A., Dagg, M.J., Whitledge, T.E.,
For information on how to obtain the data used to produce the Dortch, Q., 1999. Nutrients, irradiance, and mixing as factors regulating primary
results of this paper, contact Hongyi Yao (Jason.a.yao@hotmail. production in coastal waters impacted by the Mississippi River plume. Cont.
Shelf Res. 19, 1113–1141.
com). This research was supported by National Natural Science Masse, A.K., Murthy, C.R., 1992. Analysis of the Niagara river plume dynamics. J.
Foundation of China (NSFC) Grant no. 51479074 and the ACS-PRF Geophys. Res.-Oceans 97, 2403–2420.
Program Award 51128-ND8. McKee, B.A., Nittrouer, C.A., DeMaster, D.J., 1983. Concepts of sediment deposition
and accumulation applied to the continental shelf near the mouth of the
Yangtze River. Geology 11, 631–633.
Milliman, J.D., Shen, H.T., Yang, Z.S., Mead, R.H., 1985a. Transport and deposition of
river sediment in the Changjiang estuary and adjacent continental shelf. Cont.
References Shelf Res. 4, 37–45.
Milliman, J.D., Shen, H.T., Yang, Z.S., Meade, R.H., 1985b. Transport and deposition of
Allison, M.A., Lee, M.T., Ogston, A.S., Aller, R.C., 2000. Origin of Amazon mudbanks river sediment in the changjiang estuary and adjacent continental-shelf. Cont.
along the northeastern coast of South America. Mar. Geol. 163, 241–256. Shelf Res. 4, 37–45.
Anthony, E.J., Gardel, A., Gratiot, N., Proisy, C., Allison, M.A., Dolique, F., Fromard, F., Nardin, W., Fagherazzi, S., 2012. The effect of wind waves on the development of
2010. The Amazon-influenced muddy coast of South America A review of mud- river mouth bars. Geophys. Res. Lett. 39, 12.
bank-shoreline interactions. Earth-Sci. Rev. 103, 99–121. Nardin, W., Mariotti, G., Edmonds, D.A., Guercio, R., Fagherazzi, S., 2013. Growth of
Avicola, G., Huq, P., 2003a. The characteristics of the recirculating bulge region in river mouth bars in sheltered bays in the presence of frontal waves. J. Geophys.
coastal buoyant outflows. J. Mar. Res. 61, 435–463. Res.: Earth Surf. 118 (2), 872–886.
Avicola, G., Huq, P., 2003b. The role of outflow geometry in the formation of the Nof, D., Pichevin, T., 2001. The ballooning of outflows. J. Phys. Oceanogr. 31,
recirculating bulge region in coastal buoyant outflows. J. Mar. Res. 61, 411–434. 3045–3058.
Beardsley, R.C., Hart, J., 1978. A simple theoretical model for the flow of an estuary Odonnell, J., 1990. The formation and fate of a river plume - a numerical-model. J.
onto a continental shelf. J. Geophysi. Res. 83, 873–883. Phys. Oceanogr. 20, 551–569.
Brown, W.M., Ritter, J.R., 1971. Sediment Transport and Turbidity in the Eel River Pichevin, T., Nof, D., Lutjeharms, J., 1999. Why are there Agulhas rings? J. Phys.
Basin. Center for Integrated Data Analytics Wisconsin Science Center, California. Oceanogr. 29, 693–707.
Chao, S.Y., Boicourt, W.C., 1986. Onset of estuarine plumes. J. Phys. Oceanogr. 16, Pimenta, F.M., Kirwan, A.D., Huq, P., 2011. On the transport of buoyant coastal
2137–2149. plumes. J. Phys. Oceanogr. 41, 620–640.
Chen, Z., Song, B., Wang, Z., Cai, Y., 2000. Late Quaternary evolution of the sub- Shapiro, G.I., 2004. A 2.5D model for sand transport in a shallow sea: effect of
aqueous Yangtze Delta, China: sedimentation, stratigraphy, palynology, and Ekman veering. Cont. Shelf Res. 24, 659–671.
deformation. Mar. Geol. 162, 423–441. Sommerfield, C.K., Nittrouer, C.A., 1999. Modern accumulation rates and a sediment
Fagherazzi, S., Edmonds, D.A., Nardin, W., Leonardi, N., Canestrelli, A., Falcini, F., budget for the Eel shelf: a flood-dominated depositional environment. Mar.
Jerolmack, D.J., Mariotti, G., Rowland, J.C., Slingerland, R.L., 2015. Dynamics of Geol. 154, 227–241 215.
river mouth deposits. Rev. Geophys. 53, 642–672. http://dx.doi.org/10.1002/ Traykovski, P., Geyer, W.R., Irish, J.D., Lynch, J.F., 2000. The role of wave-induced
2014RG000451. density-driven fluid mud flows for cross-shelf transport on the Eel River con-
Falcini, F., Fagherazzi, S., Jerolmack, D.J., 2012a. Wave-supported sediment gravity tinental shelf. Cont. Shelf Res. 20, 2113–2140.
flows currents: Effects of fluid-induced pressure gradients and flow width Van Leussen, W., 1988. Aggregation of Particles, Settling Velocity of Mud Flocs a
spreading. Cont. Shelf Res. 33, 37–50. Review, Physical Processes in Estuaries. Springer, pp. 347–403.
Falcini, F., Khan, N.S., Macelloni, L., Horton, B.P., Lutken, C.B., Mckee, K.L., Santoleri, Warne, A.G., Meade, R.H., White, W.A., Guevara, E.H., Gibeaut, J., Smyth, R.C., Aslan,
R., Colella, S., Li, C., Volpe, G., 2012b. Linking the historic 2011 Mississippi River A., Tremblay, T., 2002. Regional controls on geomorphology, hydrology, and
flood to coastal wetland sedimentation. Nat. Geosci. 5, 803–807. ecosystem integrity in the Orinoco Delta, Venezuela. Geomorphology 44,
Fong, D.A., Geyer, W.R., 2002. The alongshore transport of freshwater in a surface- 273–307.
trapped river plume. J. Phys. Oceanogr. 32, 957–972. Williams, G.P., 1989. Sediment concentration versus water discharge during single
Garvine, R.W., 1987. Estuary plumes and fronts in shelf waters - a layer model. J. hydrologic events in rivers. J. Hydrol. 111, 89–106.
Phys. Oceanogr. 17, 1877–1896. Winterwerp, J.C., 2002. On the flocculation and settling velocity of estuarine mud.
Garvine, R.W., 1995. A dynamical system for classifying buoyant coastal discharges. Cont. Shelf Res. 22, 1339–1360.
Cont. Shelf Res. 15, 1585. Wiseman, W.J., Garvine, R.W., 1995. Plumes and coastal currents near large river
Garvine, R.W., 1999. Penetration of buoyant coastal discharge onto the continental mouths. Estuaries 18, 509–517.
shelf: a numerical model experiment. J. Phys. Oceanogr. 29, 1892–1909. Wright, L.D., 1977. Sediment transport and deposition at river mouths - synthesis.
Geyer, W.R., Hill, P.S., Kineke, G.C., 2004a. The transport, transformation and dis- Geol. Soc. Am. Bull. 88, 857–868.
persal of sediment by buoyant coastal flows. Cont. Shelf Res. 24, 927–949. Wright, L.D., Friedrichs, C.T., Kim, S.C., Scully, M.E., 2001. Effects of ambient currents
Geyer, W.R., Signell, R.P., Fong, D.A., Wang, J., Anderson, D.M., Keafer, B.A., 2004b. and waves on gravity-driven sediment transport on continental shelves. Mar.
The freshwater transport and dynamics of the western Maine coastal current. Geol. 175, 25–45.
Cont. Shelf Res. 24, 1339–1357. Yankovsky, A.E., Chapman, D.C., 1997. A simple theory for the fate of buoyant
Gibbs, R.J., 1985. Estuarine flocs - their size, settling velocity and density. J. Geo- coastal discharges. J. Phys. Oceanogr. 27, 1386–1401.
phys. Res.-Oceans 90, 3249–3251. Yankovsky, A.E., Hickey, B.M., Munchow, A.K., 2001. Impact of variable inflow on
Hetland, R.D., Hsu, T.J., 2013. Freshwater and sediment dispersal in large river the dynamics of a coastal buoyant plume. J. Geophys. Res.-Oceans 106,
plumes. Biogeochemical Dynamics at Large River-Coastal Interfaces: Linkages 19809–19824.
with Global Climate Change, 55–85. You, Z.J., 2004. The effect of suspended sediment concentration on the settling
Hickey, B.M., Pietrafesa, L.J., Jay, D.A., Boicourt, W.C., 1998. The Columbia River velocity of cohesive sediment in quiescent water. Ocean Eng. 31, 1955–1965.
Plume Study: subtidal variability in the velocity and salinity fields. J. Geophys. Yu, D., 2006. Analysis on Impact of the Coriolis Forces on Flow in the Yangtze River
Res.-Oceans 103, 10339–10368. Estuary, Third Chinese-German Joint Symposium on Coastal and Ocean En-
Horner-Devine, A.R., Fong, D.A., Monismith, S.G., Maxworthy, T., 2006. Laboratory gineering. National Cheng Kung University, Tainan.

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy