![Cross channel contour plots of (a) along channel velocity, (b) cross channel velocity, (c) density, (d) log10 of Richardson ]normalized by 0.25, and (e) total suspended solids at hour 0.](https://www.researchgate.net/profile/Lawrence-Sanford/publication/223292982/figure/fig2/AS:956802065784833@1605130797446/Cross-channel-contour-plots-of-a-along-channel-velocity-b-cross-channel-velocity_Q320.jpg)
Content uploaded by Lawrence P. Sanford
Author content
All content in this area was uploaded by Lawrence P. Sanford on Nov 11, 2020
Content may be subject to copyright.
Continental Shelf Research 27 (2007) 679–698
Lateral dynamics and associated transport of sediment
in the upper reaches of a partially mixed estuary,
Chesapeake Bay, USA
David C. Fugate
a,
, Carl T. Friedrichs
b
, Lawrence P. Sanford
c
a
Department of Marine and Environmental Science, Florida Gulf Coast University, 10501 FGCU Blvd. S., FL, USA
b
Virginia Institute of Marine Science, College of William and Mary, PO Box 1346, Gloucester Point, VA 23062, USA
c
Unversity of Maryland Center for Environmental Science, Horn Point Laboratory, PO Box 775, Cambridge, MD 21613, USA
Received 6 September 2005; received in revised form 27 October 2006; accepted 7 November 2006
Available online 10 January 2007
Abstract
Data from time series of transects made over a tidal period across a section of the upper Chesapeake Bay, USA, reveal
the influence of lateral dynamics on sediment transport in an area with a deep channel and broad extents of shallower
flanks. Contributions to lateral momentum by rotation (Coriolis plus channel curvature), cross channel density gradients
and cross channel surface slope were estimated, and the friction and acceleration terms needed to complete the balance
were compared to patterns of observed lateral circulation. During ebb, net rotation effects were larger because of river
velocity and reinforcement of Coriolis by curvature. During flood, stratification was greater because of landward advection
of strong vertical density gradients. Together, the ebb intensified lateral circulation and flood intensified stratification
focused sediment and sediment transport along the left side of the estuary (looking seaward). The tendency for greater
stratification on flood and net sediment flux toward the left-hand shoal are contrary to more common models which, in the
northern hemisphere, predict greater resuspension on flood and move sediment toward the right-hand shoal. These tidal
asymmetries interact with the lateral circulation to focus net sediment flux on the left side of the estuary, and to produce
net ebb directed sediment transport at the surface of the same order of magnitude as net flood directed sediment transport
at the bottom.
r2006 Elsevier Ltd. All rights reserved.
Keywords: Sediment transport; Sediment dynamics; Transverse circulation; Momentum balance; Estuarine dynamics;
USA Chesapeake Bay
1. Introduction
Although numerous studies have investigated the
dynamics of lateral circulation in partially mixed
estuaries, few have addressed the effect of lateral
circulation on the net transport of suspended sediment
and other buoyant material. This paper will show that
lateral circulation has a significant effect on the cross
channel distribution of the suspended sediment load,
and consequently, the distribution of net flux of
sediment across the estuary.
The principal forcings that drive transverse circula-
tion in estuaries are transverse density gradients,
ARTICLE IN PRESS
www.elsevier.com/locate/csr
0278-4343/$ - see front matter r2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.csr.2006.11.012
Corresponding author. Tel.: +01 239 590 7136;
fax: +01 239 590 7200.
E-mail addresses: dfugate@fgcu.edu (D.C. Fugate),
cfried@vims.edu (C.T. Friedrichs),
lsanford@hpl.umces.edu (L.P. Sanford).
Coriolis acceleration of the along channel current, and
centrifugal accelerations from flows around bends. Li
and Valle-Levinson (1999) have also suggested that
lateral currents may be generated from tidal flows
interacting with channel bathymetry independent of
density gradients. Differential advection between
shoals and channels of the along channel density
gradient produces an across-channel density gradient
(Huzzey and Brubaker, 1988;Nunes Vaz and
Simpson, 1985;Ridd et al., 1998;Swift et al., 1996).
At the end of flood in a well-mixed estuary, faster
currents over the deep channel will have emplaced
denser water over the channel than over the flanks,
resulting in counter rotating circulation cells that
converge at the surface over the channel. Baroclinic
forced transverse flow may be reduced by vertical
density gradients (Geyer et al., 1998;Lerczak and
Geyer, 2004;Valle-Levinson et al., 2003a)orby
turbulent mixing in the deep channel (Lacy et al.,
2003). Friedrichs and Valle-Levinson (1998) and Mied
et al. (2002) have shown that transverse circulation
may also be explained by phase lags between channels
and shoals and concomitant convergence and diver-
gence of Coriolis forcing. Cross channel flows driven
by Coriolis forcing are predicted to be strongest in
large estuaries at the end of ebb when flood has
already started in the shoals, and to be centered over
the slope break. Coriolis forcing and centrifugal
accelerations are dynamically similar (Chant, 2002;
Geyer, 1993;Hench et al., 2002)underasteadyflow.
Faster surface flows have higher Coriolis or centrifu-
gal forcing than near bottom flows and set up a single
rotating cell across the estuary. The interaction
between time varying stratification in the water
column and these major forcings has a complex effect
on the lateral circulation (Lerczak and Geyer, 2004)
and as shown in this paper, also on the cross channel
transport of suspended sediment.
Geyer et al. (1998) investigated the effect of
lateral circulation on sediment transport and focus-
ing in the ETM region of the Hudson River.
Centrifugal acceleration, Coriolis forcing, and setup
of transverse baroclinic pressure gradients by
differential advection of along-channel density
gradients drove the transverse flow. During flood,
these accelerations were in the same direction,
creating a transverse flow across the breadth of
the estuary that trapped sediment on the western
shoal. During ebb, the Coriolis and baroclinic
pressure gradients were in opposition; consequently
transverse currents were smaller and failed to focus
sediment onto the opposite bank.
Along channel flow and the cross channel
distribution of sediment concentration determines
the distribution of net sediment flux across a section
of an estuary. Pritchard (1967) characterized estu-
aries by their residual estuarine circulation. Verti-
cally homogenous well-mixed estuaries are ‘vertical’
estuaries, a reference to the region of no residual
motion that is an approximately vertical line
between the surface and the bottom of the channel.
Coriolis forcing results in net flow up the estuary on
the right side (looking upstream in the Northern
hemisphere) and down the estuary on the left. In
contrast, he characterized partially mixed and well-
stratified estuaries as ‘horizontal’ estuaries because
the region of no net motion is an approximately
horizontal line across the mid-water column. Net
flow at the surface is out of the estuary and net flow
at the bottom is into the estuary. Tidal circulation in
partially mixed estuaries such as Chesapeake Bay
often results in strained induced periodic stratifica-
tion (Simpson et al., 1990) in which relatively well-
mixed flood currents are followed by more stratified
ebb currents, producing a mid-depth region of no
net motion. However, near the upper extent of the
salinity intrusion in partially mixed estuaries, the
along-channel circulation deviates from this pattern.
This paper will discuss the mean and time
evolution of cross channel momentum balance,
cross channel circulation, and cross channel dis-
tribution of suspended sediment transport near the
ETM region of the Chesapeake Bay. The impor-
tance of shoals to creating differential advection and
lateral circulation has already been established in
the literature. Here we will show also that patterns
of mixing, resuspension and lateral advection over
the shoals are an important component of the cross
channel distribution of net along channel sediment
transport. Additionally, we will show that regions of
axially varying stratification, which are especially
common in the upper extent of estuarine salinity
intrusions, include areas where along-channel ad-
vection causes the water column to be more
stratified during flood than ebb, and this impacts
cross channel sediment transport and net along
channel sediment flux.
2. Site and methods
The Chesapeake Bay is a large estuary with a
length of 300 km and width of up to 30 km located
on the mid Atlantic bight of the continental USA
(Fig. 1). The northern region of the Chesapeake Bay
ARTICLE IN PRESS
D.C. Fugate et al. / Continental Shelf Research 27 (2007) 679–698680
near the study site is relatively straight and typical
of partially mixed estuaries in the Northern Hemi-
sphere in that it has a deeper main channel on
the left looking seaward and wide shoal areas on the
right. At the cross-section shown in Fig. 1, the
eastern flank (left side looking seaward) is narrower
with a relatively steep incline to the 14 m deep main
channel. The slope from the channel to the western
flank is gradual and the western flank is much
broader than the eastern flank. The mean tidal
range is about 0.5 m in this section of the Bay.
Although the turbidity maximum is often found in
the general area of the northern Chesapeake Bay, it
was not located at this exact study site when the
measurements for the present study were made.
Nevertheless, the results from this study should bear
upon lateral variations of sediment flux in the ETM.
The RV Elis Olsson repeated transects about
every 50 min from station E on the east side of the
Bay to station A on the west side on October 9, 2002
over a spring tidal cycle for a total of 14 transects
(Fig. 1). Velocity and acoustic backscatter ampli-
tudes were collected from a bow mounted 1200 kHz
RDI acoustic doppler current profiler (ADCP)
sampling at a frequency of 0.5 Hz and vertical
resolution of 0.25 m. The research vessel traveled at
a constant speed of about 3 m s
1
across the
2700 m transect, so that individual profiles taken
by the ADCP are about 6 m apart. On the return
trip, profiles were taken at each of 5 stations with a
ARTICLE IN PRESS
76.70 76.50 76.30 76.10 75.90
38.90
39.10
39.30
39.50
Baltimore
Havre de Grace
E
A
Transect
Longitude W
Latitude
Fig. 1. Study site in the Northern Chesapeake Bay showing the transect location, the westernmost station A, and the easternmost station
E. Shading indicates depth contours in increments of 5 m.
D.C. Fugate et al. / Continental Shelf Research 27 (2007) 679–698 681
CTD and a LISST particle size analyzer. The
westernmost station is station A, and the other 4
stations were located 900, 1500, 2000, and 2700 m
east of station A. Pumped water samples were also
collected at the stations to calibrate the ADCP
backscatter for total suspended solids (TSS). The
resulting linear calibration had r
2
¼0.77 between
TSS and backscatter. This sampling regime pro-
duced a good temporal and spatial resolution of the
tidal variation of cross channel velocities, density
and suspended sediment concentrations throughout
an entire tidal cycle.
On the same day that the RV Elis Olsson collected
lateral transects, the RV Cape Henlopen repeated
axial transects along a 16 km segment of the
navigation channel in the same region of the upper
Chesapeake Bay. The along-channel stations were
spaced at 4 km intervals and were occupied about
once every 2 h. The data collected by the Cape
Henlopen focused on biological parameters, but
also included CTD with optical backscatter cali-
brated for TSS (Sanford et al., 2005). Much longer
(70 km) axial CTD transects were collected by the
Cape Henlopen on October 8 and October 11,
centered on ebb tide.
Tidally averaged quantities such as net sediment
flux and net water volume flux were calculated from
the lateral transect data by interpolating the data
from each of the 14 transects onto a uniform
14 2750 m grid with a vertical resolution of 0.25 m
and a horizontal resolution of 10 m. In order to
avoid error produced by side lobe effects of the
ADCP near the bottom, only bins above 1.75 mab
(meters above the bed) were used in the calculations.
Velocity and suspended sediment for the top meter
of the water column were linearly extrapolated from
the 4 bins in the interval of 1 and 2 m below the
surface. Any negative values of suspended sediment
resulting from this extrapolation were set to 0. Net
fluxes and residual velocities were smoothed by a
40 m (4 point) moving median in the horizontal and
a 1 m (4 point) moving median in the vertical.
Smoothing was necessary to reduce the random
noise in the raw ADCP data and does not affect the
interpretation of the results. Velocity data were
rotated across the section so that there was
minimum variance in the cross channel directed
current (positive xand ulandward to the north,
positive yand vto the west). In the bottom 2 m of
the deep channel cross channel density gradients
are set to the value measured at the depth of the
second deepest profile. Richardson numbers, Ri, are
calculated from the density and along channel
velocity grids and defined as
Ri ¼g
r
dr
dz=du
dz
2
,
where gis gravity, ris density, and uis the along
channel velocity. Richardson numbers were calcu-
lated from vertical intervals of 1 m. Normalized
Richardson numbers are presented in the figures
and are calculated as Ri/0.25.
The Coriolis and density gradient terms in the
across-channel momentum equation were then
estimated from observations for a vertical profile
near the middle of each cross-section. The form of
the across-channel momentum equation considered
is as follows:
Dv
Dtþfu u2
Rþg@Z
@yþðzÞg
r
@r
@y@
@zAz
@v
@z
¼0,
where tis time, fis the Coriolis parameter, Ris the
radius and direction of channel curvature, gis
gravity, Zis sea surface elevation, ris density, A
z
is
eddy viscosity, and local and advective acceleration
components (other than channel curvature) have
been combined into one term (Dv/Dt).
A useful way to sort the various accelerations and
forces in the across-channel momentum equation is
to separate the major components due to along-
channel velocity from those that involve across-
channel velocity. For channelized flow at this scale,
the along-channel flow contributes to across-chan-
nel momentum mainly through the Coriolis term
and through channel curvature effects in advective
acceleration. For uniform flow around bends, the
contribution of flow curvature to advective accel-
eration is approximately u
2
/R, where Ris the radius
of curvature of the channel. Based on Fig. 1,Rat
the sampling site is on the order of 40 km. For an
along-channel current of 60 cm s
1
this results in a
centrifugal acceleration that is about 1/10 of
Coriolis at the latitude of the Chesapeake Bay. At
this site during flood tide the flow curvature term
partly cancels Coriolis, while during ebb it enhances
Coriolis. Since the form (including depth-depen-
dence) and externally imposed nature of these two
terms are so similar, we will group them together as
‘‘rotation’’ ¼fuu
2
/Rin the following analysis.
If the contributions of across-channel velocity to
the momentum equation were negligible and there
were no measurement errors, then the opposite of
the sum of the observed rotation and density
ARTICLE IN PRESS
D.C. Fugate et al. / Continental Shelf Research 27 (2007) 679–698682
gradient terms would be constant in zand exactly
equal to the term associated with sea surface slope.
We therefore define the depth-averaged sum of these
terms to be the ‘‘estimated surface slope’’ in order to
provide insight into the likely sign and magnitude of
the cross-channel barotropic pressure gradient. This
approach does not assume that the remaining
acceleration and friction terms are zero. Rather, it
only assumes that the sum of the remaining terms
and the contribution of our observational error
have a small depth-averaged component.
The remaining depth-varying residual represents
imbalances between the three other lateral momen-
tum terms that, if they were in balance, would
require no lateral circulation. This residual contains
the only components of across-channel momentum
explicitly involving v, namely qv/qtand friction. It is
a bit difficult to fully interpret because it is the
combination of acceleration and frictional terms, so
it does not indicate the magnitude or sign of either
one alone. However, the net depth-varying residual
does provide an indication of the net forces acting in
the lateral direction, so provides useful insight into
the cross-channel circulation pattern. The data do
not support further decomposition of this term. We
label this depth-varying residual the ‘‘estimated qv/
qt+friction’’.
To summarize, the components of the across-
channel momentum equation estimated from ob-
served along-channel velocity (u) and observed
density (r) are calculated as follows:
Rotation ¼fuu
2
/R, where f¼0.92 10
4
s
1
and R¼40 km.
Density gradient ¼(z)(g/r)dr/dy.
Surface slope ¼(1) Depth average of
(Rotation+Density gradient),
dv/dt+Friction ¼(1) (Rotation+Density
gradient+Surface slope).
3. Results and discussion
3.1. General physical environment
The study site area is a region of interesting
hydrodynamics near the northern extent of the
salinity intrusion into the Chesapeake Bay. It is a
transition zone with a moderate energy regime and
moderate turbidity and biological activity. As such,
it typifies many upstream zones of estuaries on the
Atlantic coast of the USA, and other passive
margins (Sanford et al., 2005). The estuarine
turbidity maximum migrates through this region,
depending upon freshwater flow from the Susque-
hanna River, meteorological conditions and other
factors. Freshwater flow was generally low during
2002 and during the survey period resulted in a
region of strong vertical stratification near the study
site (Fig. 2). Fig. 2a and b show longitudinal
sections of salinity and turbidity contours of the
entire upper Chesapeake Bay made during ebb 1
day before and 2 days after the lateral transects
were made. Maximum vertical stratification is
consistently located a few km seaward of river
25 km (Fig. 2a, b), which is the location of the
lateral transects discussed in this paper. Shorter
longitudinal transects made during the day of this
survey show that advection of the horizontal
gradient in vertical stratification produced stronger
stratification at the end of flood tide at 25 km
(Fig. 2c), and relatively more mixed conditions at
the end of ebb tide (Fig. 2d). Peak along channel
currents over the channel reached 0.65 m s
1
, while
peak currents over the flanks reached about
0.4 m s
1
. Salinity ranged from 5 to 10 ppt, and
ranged from stratified to well mixed at different
locations along the transect and over the tidal
period. Suspended sediment concentrations went up
to 40 mg l
1
. Mean wind speeds were about
3.5 m s
1
, and rainfall during and before the cruise
was minimal.
3.2. Time evolution of physical environment
The first lateral transect (hour 0) occurs near the
beginning of flood tide (Fig. 3). Along channel
velocities at the bottom of the channel and the
eastern shoal were flooding at about 0.5 m s
1
. The
western flank velocities are flooding about
0.25 m s
1
, while surface currents in the channel
are still ebbing at about 0.1 m s
1
. The water
column is stratified with density contours tilted
upwards towards the eastern bank. Lateral currents
are generally very weak. Rotation and baroclinic
forcing both increase with depth and nearly balance,
completing a geostrophic balance without needing
much surface slope or across-channel current
(Fig. 4a). It is worth noting that there must be
some small surface slope associated with the
baroclinic pressure gradient, however, it does not
contribute significantly to the dynamics. Suspended
sediment (TSS) concentrations are generally near
background (10 mg l
1
) except at the channel
bottom and the western flanks where concentrations
are over 20 mg l
1
. Although the overall current
ARTICLE IN PRESS
D.C. Fugate et al. / Continental Shelf Research 27 (2007) 679–698 683
magnitudes are about the same in both flanks,
stratification on the eastern flank associated with
the tilted isopycnals dampens resuspension there.
By hour 2.0, along channel flood currents have
increased to over 0.6 m s
1
in the middle of the
water column (Fig. 5a). The greater flood current
has increased the Coriolis component of rotation,
and the estimated surface slope now tilts up more
strongly to the east (Fig. 4b). As a result, there is
now a significant depth-varying imbalance among
ARTICLE IN PRESS
Fig. 2. Longitudinal transect of the upper Chesapeake Bay from ebb tide on (a) October 8 and (b) October 11, 2002 showing salinity
contours and suspended sediment concentration (TSS). The site of the lateral transects for this study were near river 25 km. Longitudinal
transect of salinity and TSS on October 9, 2002 during (c) flood, and (d) ebb.
D.C. Fugate et al. / Continental Shelf Research 27 (2007) 679–698684
the rotation, density gradient and surface slope
terms. Cross channel baroclinic forcing reaches a
westward maximum of 4 10
5
ms
2
at about 9 m
depth while maximum rotation forcing is about
510
5
ms
2
eastward at 4 m depth (Fig. 4b). The
depth-varying imbalance is represented by the
estimated qv/qt+friction term in Fig. 4b and is
consistent with near bed flow to the west and
surface flow to the east. The observed across-
channel velocity is generally consistent with this
residual forcing, as demonstrated by a 0.3 m s
1
eastward cross channel current across most of the
channel surface and a 0.3 m s
1
westward cross
channel current developed near the bottom of the
ARTICLE IN PRESS
F
E
W
E
Meters West of Station E
Depth (m)
Depth (m)
Depth (m)
Depth (m)
Depth (m)
EW
Fig. 3. Cross channel contour plots of (a) along channel velocity, (b) cross channel velocity, (c) density, (d) log10 of Richardson
]normalized by 0.25, and (e) total suspended solids at hour 0.
D.C. Fugate et al. / Continental Shelf Research 27 (2007) 679–698 685
east flank and slope (Fig. 5b). At the surface, across-
channel currents converge over the top of the
eastern slope where slower along channel currents
over the shoal produce the maximum lateral
convergence in rotational forcing. Near the bottom,
cross channel currents converge on the eastern side
of the deep channel.
Contours of Richardson numbers at hour 2 show
the first appearance of a vertical and cross channel
distribution of enhanced and reduced mixing that
persists through most of the flood phase (Fig. 5d).
Two horizontal layers of reduced mixing associated
with stratification cross the channel, the upper one
starting at about 2 m depth in the eastern channel,
dipping to the western break in slope on the other
side at about 6 m depth. The lower layer is just
above the deep channel and intersects with the
eastern side of the estuary at 8–10 m depth and
intersects the western side at 11–12 m depth. The
western shoal of the estuary is completely mixed low
salinity water with sub-critical Richardson numbers.
Sediment resuspension reaches the surface on the
ARTICLE IN PRESS
Fig. 4. Vertical profiles of cross channel baroclinic forcing, rotational forcing, dv/dt+friction, and estimated surface slope at six times in
the tidal cycle. All terms estimated at station D, the deepest point in the main channel.
D.C. Fugate et al. / Continental Shelf Research 27 (2007) 679–698686
western flank where stratification is low and on the
eastern side of the deep channel near where lateral
surface currents converge. The general trend toward
increasing stratification (with the exception of the
western shoal) over the course of the flood is
consistent with landward advection of the more
stratified water.
By hour 3.6 flood currents are still strong in the
water column above the deep channel but slacking
on the flanks (Fig. 6). Vertical shear in the along-
channel velocity has been reduced somewhat and
the upper pycnocline has shifted farther toward the
surface, both presumably the result of at least
some vertical mixing in the middle water column.
ARTICLE IN PRESS
F
E
W
E
Meters West of Station E
Depth (m)
Depth (m)
Depth (m)
Depth (m)
Depth (m)
EW
Fig. 5. Cross channel contour plots of (a) along channel velocity, (b) cross channel velocity, (c) density, (d) log10 of Richardson
]normalized by 0.25 and (e) total suspended solids at hour 2.0.
D.C. Fugate et al. / Continental Shelf Research 27 (2007) 679–698 687
The dense water near the bed in the eastern side has
likely been advected since hour 2 by lateral
circulation toward the western side of the estuary,
reducing the weak baroclinic forcing near the
bottom (Fig. 4c), and generally leveling the pycno-
clines. The lack of shear and flat pycnoclines over
the main channel is consistent with a mainly
barotropic and largely geostrophic lateral balance
over the main channel (Fig. 4c). The two-layer
cross-channel velocity circulation is interrupted by
vertical mixing and confinement above the bottom
pycnocline (Fig. 6). This effect of localized mixing in
the main channel is indicated by a vertical plume
evident in the ADCP backscatter (Fig. 6e) located
ARTICLE IN PRESS
F
E
W
E
Meters West of Station E
Depth (m)
Depth (m)
Depth (m)
Depth (m) Depth (m)
EW
Fig. 6. Cross channel contour plots of (a) along channel velocity, (b) cross channel velocity, (c) density, (d) log10 of Richardson
]normalized by 0.25 and (e) total suspended solids at hour 3.6.
D.C. Fugate et al. / Continental Shelf Research 27 (2007) 679–698688
where the near bottom lateral currents converge
above the bottom pycnocline. The Richardson
number suggests less mixing has occurred on the
eastern shoal relative to the western shoal. While the
more stratified eastern flank is relatively devoid of
suspended sediment, the well-mixed western flank
has considerable resuspension of sediment. During
the next few hours along and cross channel
velocities subside and suspended sediment concen-
trations drop to background levels. The pycnocline
continues to relax, resulting in cross channel density
gradients near zero. In general, stratification asso-
ciated with along-channel advection continues to
grow towards the beginning of ebb around hour 5.2
(not shown).
Ebb currents begin in the flanks first and by hour
7.5 cross channel circulation has set up in the
opposite direction than during flood (Fig. 7). The
rotation term has now switched sign (with Coriolis
and curvature now additive), but the pycnocline
remains relatively flat, resulting in a relatively small
baroclinic term (Fig. 4d). The sea surface is now
estimated to slope strongly up to the west to partly
balance rotation. Rotation and barotropic forcing
result in a near geostrophic balance, but with
an opposite direction from that during flood. The
qv/qt+friction is small but also opposite to that
during flood. The result is that near bottom
transverse currents are eastward across most of
the estuary and westward at the surface. A surface
cross channel velocity divergence is now located
above the top of the eastern slope due to a lateral
divergence in the rotational forcing. The momen-
tum balance at hour 8.6 is similar to hour 7.5
(Fig. 4e). Rotational forcing continues to dominate
baroclinic pressure gradients throughout the ebb,
although as ebb proceeds, near bottom eastward
currents push saltier water to the east and begin
to set up a weak cross channel density gradient
that opposes the near bottom current. Maximum
ebb currents are at the surface only slightly
westward of the deepest part of the channel.
Stratification breaks down in the lower water
column during ebb, likely due to a combination of
high velocity shear and downstream advection
of the more well mixed water column found
upstream (Fig. 8d).
During ebb, there is considerable resuspension
over the western bank, as during flood. But now
there is also considerable resuspension over the
eastern flank. Suspended sediment is swept up the
eastern slope by the eastward near-bottom current.
Furthermore, the previously relaxed pycnocline is
still below the edge of the eastern shoal. By hour 8.6
there is a dramatic vertical plume over the top of the
eastern slope that persists for several hours (Fig. 8e).
Note that this plume is associated with the top of
the eastern slope while the plume produced by
localized vertical mixing during flood currents is
located at the base of the eastern slope.
As ebb phase progresses, along and cross channel
currents subside. By the end of ebb, the previously
eastward near bed currents have finally advected
salty water back up onto the eastern shoal,
significantly increasing the lateral density gradient
and, in combination with along-channel advection,
enhancing stratification on the eastern shoal
(Fig. 9c). With the pycnocline raised onto the
eastern shoal, sediment suspension is reduced there
relative to the adjacent deeper area underneath the
pycnocline (Fig. 9e). At hour 11.1 the net rotation
term is weak but has opposite signs at surface and
bottom (Fig. 4f). Baroclinic forcing increases with
depth, similar to hour 0.0. The combination of
rotation and baroclinic forcing near the surface
results in a weak surface slope that is still tilting up
toward the west (Fig. 4f). The relaxation of the
rotational forcing allows a near bottom westward
baroclinic driven current to set up a rotating cell in
the eastern side of the transect that converges with
the remnant rotating cell in the western side of the
transect (Fig. 9b). Suspended sediment that was
focused by the lateral currents of the ebb tide is
mixed and retained in the eastern side by the eastern
rotating cell. The remaining qv/qt+friction terms
are consistent with the eastern rotating cell located
above the deep channel.
In summary, the momentum balance throughout
the tidal cycle appears to be primarily geostrophic.
The inferred across-channel surface slope ap-
proaches zero near slack water, is most negative
near maximum flood (when the eastern shoal has a
higher surface elevation), and is most positive
(highest absolute value of about 0.7 10
4
ms
2
/
9.81 m s
2
¼.71 10
5
) near maximum ebb. Over
3 km of channel width, this corresponds to a lateral
height difference of about 2 cm, which is reasonable.
The baroclinic term builds during ebb, reaching its
maximum magnitude at the end of ebb/beginning of
flood and dissipating during flood. The rotation
term is strong during both flood and ebb, but
slightly stronger during ebb, in part because channel
curvature reinforces Coriolis here during ebb. Since
the radius of curvature is relatively large (40 km),
ARTICLE IN PRESS
D.C. Fugate et al. / Continental Shelf Research 27 (2007) 679–698 689
it is still appropriate to describe the main balance as
geostrophic.
However, there are sufficient imbalances to drive
a significant time-varying lateral circulation. This
circulation has two components. The first is
essentially a time-varying adjustment to changing
near-geostrophic structure over the tidal cycle. For
example, assuming qv/qtoV, where ois the tidal
frequency, then the magnitude of Vbased on the
indirect measure of qv/qtin the surface layer in
Fig. 4b should be about (0.2 10
4
ms
2
)/
(1.4 10
4
s
1
)¼0.14 m s
1
, which is similar to
ARTICLE IN PRESS
F
E
W
E
Meters West of Station E
Depth (m)
Depth (m)
Depth (m)
Depth (m)
Depth (m)
EW
Fig. 7. Cross channel contour plots of (a) along channel velocity, (b) cross channel velocity, (c) density, (d) log10 of Richardson
]normalized by 0.25 and (e) total suspended solids at hour 7.5.
D.C. Fugate et al. / Continental Shelf Research 27 (2007) 679–698690
surface layer lateral velocities in Fig. 5. The second
is frictional dissipation of lateral pressure gradients
in the bottom Ekman layer, as evidenced by
generally westward near-bottom currents on flood
and eastward near-bottom currents on ebb. Because
the bottom 1.5 m of the water column is not
resolved by the ADCP, some of this bottom Ekman
layer is not visible in the transect data.
Overall, stratification tends to increase over the
course of the flood, which is consistent with
landward advection of more stratified water from
downstream. The eastern shoal is generally more
ARTICLE IN PRESS
F
E
W
E
Meters West of Station E
Depth (m)
Depth (m)
Depth (m)
Depth (m)
Depth (m)
EW
Fig. 8. Cross channel contour plots of (a) along channel velocity, (b) cross channel velocity, (c) density, (d) log10 of Richardson
]normalized by 0.25 and (e) total suspended solids at hour 8.6.
D.C. Fugate et al. / Continental Shelf Research 27 (2007) 679–698 691
stratified than the western shoal and is subject to
less intense sediment resuspension. Convergence of
cross-channel currents occurs more often on the
eastern side, which presumably leads to sediment
accumulation on the eastern side under the condi-
tions observed during this tidal cycle.
3.3. Net transport patterns
The residual along channel velocities reflect the
across-channel variation in estuarine dynamics in
this upper extent of the estuary’s salt intrusion
(Fig. 10a). A contour of the level of no net motion
ARTICLE IN PRESS
F
E
W
E
Meters West of Station E
Depth (m)
Depth (m)
Depth (m)
Depth (m)
Depth (m)
EW
Fig. 9. Cross channel contour plots of (a) along channel velocity, (b) cross channel velocity, (c) density, (d) log10 of Richardson
]normalized by 0.25 and (e) total suspended solids at hour 11.1.
D.C. Fugate et al. / Continental Shelf Research 27 (2007) 679–698692
describes a distorted sideways S shape that follows
the bathymetry on the east side but intersects almost
vertically with the bottom west of the deep channel.
Two centers of mean seaward flow separated by a
near surface landward directed core is typical of
mean along-channel circulation in a weakly strati-
fied estuary over a triangular shaped channel
(Friedrichs and Hamrick, 1996;Wong, 1994). Here
the stronger surface flow is shifted to the west with
respect to the deep channel in response to combined
effects of Coriolis and channel curvature. The
resulting landward flow on the eastern shoal and
two-layered flow over the eastern flank favor saltier,
more stratified water there relative to the more
strongly seaward and one-layered flow over the
western shoal and flank (Fig. 11). This pattern is
similar to that proposed by Valle-Levinson et al.
(2003b) for cases where rotation is slightly greater
than friction effects.
The mean lateral circulation (Fig. 10b) is qualita-
tively consistent with the mean along-channel
circulation and also with greater density and
stratification toward the eastern side of the cross-
section. The tendency for westward flow near the
surface and eastward flow near the bed would
indeed tend to advect the residual seaward surface
flow toward the western shoal and advect the
landward bottom flow toward the eastern shoal.
This net pattern for observed lateral circulation is
consistent with greater rotationally induced lateral
circulation on ebb than on flood. There are three
likely contributions to greater rotational effects on
ebb. First, Coriolis and curvature affects are
additive on ebb but partially cancel on flood.
Second, riverine discharge from the Susquehanna
causes overall velocities to be stronger on ebb than
on flood, accentuating the asymmetry. And lastly,
the along channel advection of strong vertical
density gradients during flood tide produces greater
stratification during flood and dampens the lateral
circulation.
The tidally averaged density field (Fig. 11) does
not capture some important tidal asymmetries in
stratification that influence net sediment transport.
Most significantly, the east side tends to remain
stratified during much of flood, which is the main
time when westward near-bed lateral currents would
otherwise be advecting sediment off the eastern
ARTICLE IN PRESS
a
bW
E
Meters West of Station E
Depth (m)
EW
F
E
Depth (m)
Fig. 10. Cross channel contour plots of (a) along channel residual velocity (black contour line shows region of no net motion) and (b)
lateral residual velocity.
D.C. Fugate et al. / Continental Shelf Research 27 (2007) 679–698 693
shoal. More persistent stratification on the eastern
shoal during flood is in some respects counter-
intuitive, since tidal straining and bottom mixing
would normally weaken stratification during flood.
However, in this case axial advection of the salt
front preferentially strengthens near-bottom
stratification on the eastern shoal, inhibiting resus-
pension. Then during ebb, stronger shear in the
along-channel velocity and the advection of fresher,
mixed water from upstream resuspends sediment
(see Fig. 2 and hour 8.6, Fig. 8). The western shoal
is more symmetrically well mixed between flood and
ebb. Even when the salty water is advected to the
western side during flood, it quickly becomes mixed
over the western shoal (see hour 3.6, Fig. 6). Then
during ebb, the pycnocline is advected away from
the western shoal, maintaining fresher well mixed
water on the west side. Symmetrically well-mixed
conditions on the western shoal lead to high tidal
resuspension there, but little net axial sediment
transport (see below).
Given the frequent observations of the density
and along-channel velocity fields, the tidally aver-
aged values of the observed lateral momentum
terms can be interpreted in a manner similar to that
pursued for the instantaneous tidal dynamics
(Fig. 12). The substantial vertical shear in the mean
along-channel velocity (from 410 cm s
1
landward
near the bed to 410 cm s
1
seaward near the
surface) creates a significant mean rotation term
(fou4+ou
2
4/R), with a positive sign near
the bed and a negative sign near the surface. Near
the surface, the mean baroclinic forcing is small and
the mean rotation term is nearly balanced by the
surface slope. Near the bottom however, a negative
density gradient persists through most of ebb and
flood, and the mean density gradient is nearly
balanced by a combination of mean surface slope
and mean rotation. The momentum imbalance
represented by qv/qt+friction is eastward directed
near the bottom and westward directed in the
surface layer. In its tidally averaged manifestation,
this term must be primarily frictional because qv/
q00. It drives a mean lateral current over the deep
channel and across most of the estuary that is
eastward directed near the bottom and westward
directed near the surface.
There are thus two relatively large-scale effects
which result in net transport of rapidly settling
sediment toward the eastern shoal: (i) the mean
cross-channel circulation converges on the east side
near the bottom and (ii) the eastern shoal tends to
remain relatively stratified during the flood, redu-
cing resuspension during the one part of the tidal
cycle when near bed lateral currents tend to be
westward. It is worth noting that our inference that
deposition is occurring on the eastern shoal is at
odds with the relatively steeper topography on this
shoal. Morphodynamically, over the long term one
would expect net deposition on the eastern side to
cause the eastern shoal to be wider than the western
shoal. It might be that the local along-channel
gradient in stratification that ultimately drives this
process is not permanently fixed in space at this
location, so this process has not been predominantly
acting at this one location in the long run. Other
potentially confounding effects include channel
dredging and relict topography from the drowned
river valley, or the predominance of storm driven
processes to the morphology.
While mean flow near the bed moves rapidly
settling sediment to the eastern shoal, there are also
ARTICLE IN PRESS
Meters West of Station E
Depth (m)
EW
Fig. 11. Tidally averaged density field.
D.C. Fugate et al. / Continental Shelf Research 27 (2007) 679–698694
localized lateral convergence effects which tend to
focus high suspended sediment concentrations.
Residual lateral velocities (Fig. 10b) reflect the two
sites of lateral convergences on the east side that
coincide with the focusing of sediment during flood
and ebb, respectively. The flood convergence is near
bottom, slightly east of the deep channel. The ebb
convergence is further east above the top of the
eastern slope. Fig. 13a displays the tidally averaged
lateral sediment flux, which captures aspects of both
the larger scale mean lateral flux to the east and
localized convergences. It is interesting that there
are sharp changes of slope in the bottom topogra-
phy directly under the two near-bottom conver-
gences in net lateral sediment flux on the eastern
shoal.
Convergences and tidal asymmetries in lateral
circulation focus the suspended sediment at differ-
ent locations across the estuary during flood and
ebb and determine the sites of maximum along
channel net sediment transport (Fig. 13b). While
mean lateral sediment transport resembles mean
lateral circulation, mean along channel transport
differs from mean along channel circulation.
The large ebb directed sediment flux at the surface
over the top of the eastern slope is a result of
the focusing of suspended sediment by the cross
channel currents during ebb and is about the same
size as the flood directed transport near the bottom.
Peak along-channel tidal velocities tend to occur
over the deep channel during both ebb and flood,
and peak tidally averaged TSS concentrations are
higher over the western shoal (Fig. 13c), but peak
sediment transport is shifted significantly towards
the eastern side of the channel throughout the
tidal cycle.
3.4. Limitations and generalities
This data set is admittedly limited in that it
consists of only one tidal cycle. A longer time-series
is needed to confirm that these conditions are
representative of commonly present conditions in
the upper Chesapeake Bay and elsewhere, and not
simply an anomaly. For example, times of very
strong winds and waves are likely to be associated
with higher suspended sediment concentrations and
potentially important and very different net trans-
port patterns. Thus this data set is likely to only
represent lower, albeit still commonly present,
energy conditions. Other limitations include an
inability to measure very close to the bed due to
potential side lobe interference and ringing of
ADCP backscatter near the seabed. Fig. 10 shows
that there is strong shear in the mean lateral velocity
quite near the bed, for example. The diagnostic
ARTICLE IN PRESS
Fig. 12. Tidally averaged cross channel momentum terms.
D.C. Fugate et al. / Continental Shelf Research 27 (2007) 679–698 695
momentum calculations presented in Figs. 4 and 12
also represent conditions at a single location, the
deepest point of the channel, and might be some-
what different at other locations.
Nonetheless, the main driver of the patterns seen
here, namely increasing stratification on flood
rather than ebb, can be expected whenever one
samples at a transition to decreased stratification.
This should commonly occur in the upper reaches of
many estuaries, since a reduction in stratification
almost always occurs at some point as one
approaches the transition to fresh water. This
pattern of increasing stratification on flood is also
likely to be seen just downstream of secondary
turbidity maxima associated with along channel
decreases in stratification in the central reaches of
estuaries. It is becoming increasingly recognized
that along-channel decreases in depth and/or
increases in tidal velocity in estuaries can commonly
lead to a landward decrease in stratification well
downstream of the main turbidity maximum
(e.g. Lin and Kuo, 2001). This may imply net
ARTICLE IN PRESS
Meters West of Station E
Depth (m)
a
b
c
Depth (m)
Depth (m)
W
E
EW
Fig. 13. Cross channel contour plot of (a) across-channel net sediment flux, (b) along channel net sediment flux and (c) tidally averaged
TSS concentration.
D.C. Fugate et al. / Continental Shelf Research 27 (2007) 679–698696
sediment accumulation on the left shoal (looking
seaward in the northern hemisphere) at these
locations.
4. Conclusions
The upper Chesapeake Bay provides a case study
for examining the lateral dynamics and associated
transport of fine sediment in the upper reaches of a
partially mixed estuary with commonly observed
shoal-channel geometry. A useful approach in
examining the lateral momentum balance in such a
system is to directly assess the components of
momentum that do not involve measurements of
lateral velocity, namely the pressure gradient and
rotation terms, and then assess the general consis-
tency of the residual term (‘‘estimated qv/qt+fric-
tion’’) relative to the observed lateral circulation.
By doing so, it was observed that the dominant
balance during both flood and ebb is largely
geostrophic, but that important asymmetries occur.
During ebb, net rotation effects are larger for three
reasons: (i) river discharge causes the along channel
velocity to be larger, (ii) Coriolis is reinforced by
channel curvature, and (iii) the along channel
advection of strong vertical density gradients during
flood tide produces greater stratification during
flood and dampens the lateral circulation. The
larger shear in along channel velocity during ebb
compounded by rotation cannot as easily be
balanced by geostrophy, and stronger across-chan-
nel circulation during ebb results. The tidally
varying sign and relative magnitude of the observed
lateral circulation during flood and ebb are con-
sistent with the inferred qv/qq+friction term in
momentum.
Greater mean stratification on the eastern shoal
relative to the western shoal was favored by (i) mean
lateral circulation to the east near the bed, (ii) a
consistent lateral shift in the mean two-layer along-
channel velocity, and (iii) the mean tilt of the
pycnocline. However, tidal asymmetries in stratifi-
cation on the shoals were also apparent. On the
eastern shoal, along channel advection of strong
vertical density gradients moved the pycnocline
closer to the bed on flood, damping resuspension
during the one time of the tidal cycle that the (albeit
weakened) net rotational effects might otherwise
have advected nearbed sediment westward. The
western shoal remained well mixed and above the
pycnocline during both tidal stages.
Asymmetries in lateral circulation, stratification,
and resuspension resulted in net along-channel
sediment transport being concentrated toward the
eastern side of the estuary. East of the channel axis,
greater suspension consistently occurred during ebb
when stratification was weaker, augmenting the
effect of the transverse residual current. Despite a
less prominent ebb-directed mean velocity on the
eastern side of the estuary, asymmetrically greater
suspension during ebb resulted in much stronger
down-estuary sediment transport. All these sedi-
ment transport asymmetries were less dramatic on
the western side where suspension occurred more
evenly on ebb and flood.
The canonical pattern of strain-induced periodic
stratification predicts that greater stratification is
likely on ebb tide in partially mixed estuaries.
However, in regions of axially varying stratification,
which are especially common in the upper reaches
of the estuarine salinity intrusion, strong gradients
in vertical stratification may be advected up estuary
during flood, locally producing more stratification
during flood than ebb. The subsequent tidal
asymmetries in shear-driven mixing and tidal
variations in the across-channel elevation of the
pycnocline result in tidal asymmetries in sediment
resuspension. In the location of our study in the
upper Chesapeake Bay, greater stratification is seen
on flood tide, particularly on the east side of the
estuary. The resulting tidal asymmetries in sediment
resuspension interact with the lateral circulation to
focus net sediment flux on the eastern side of the
estuary, and to produce net ebb directed sediment
transport at the surface of the same order of
magnitude as net flood directed sediment transport
at the bottom.
Acknowledgments
The data for this study were collected as part of
the BITMAX (Bio-physical Interaction in the
Turbidity MAXimum) program which was funded
by the US National Science Foundation, Division
of Ocean Sciences, Collaborative Awards OCE-
0002529 and OCE-0002543. We are grateful to the
captains and crews of the RV Elis Olsson and the
RV Cape Henlopen, and particularly to Grace
Cartwright, Elizabeth North and Steve Suttles, for
their roles in collecting and processing the field
observations presented here. VIMS Contribution
No. 2783.
ARTICLE IN PRESS
D.C. Fugate et al. / Continental Shelf Research 27 (2007) 679–698 697
References
Chant, R.J., 2002. Secondary circulation in a region of flow
curvature. Journal of Geophysical Research 107 (C9), 3131.
Friedrichs, C., Hamrick, J.M., 1996. Effects of channel geometry
on cross sectional variation in along channel velocity in
partially stratified estuaries. In: Aubrey, D.G., Friedrichs,
C.T. (Eds.), Buoyancy Effects on Coastal and Estuarine
Dynamics. American Geophysical Union, Washington, DC,
pp. 283–300.
Friedrichs, C.T., Valle-Levinson, A., 1998. Transverse circulation
associated with lateral shear in tidal estuaries.
Geyer, W.R., 1993. Three-dimensional flow around headlands.
Journal of Geophysical Research 98, 955–966.
Geyer, W.R., Signell, R.P., Kineke, G., 1998. Lateral trapping of
sediment in a partially mixed estuary. In: Scheffers, D. (Ed.),
Physics of Estuaries and Coastal Seas. Balkema, Rotterdam,
pp. 115–124.
Hench, J.L., Blanton, B.O., Luettich, R.A.J., 2002. Lateral
dynamic analysis and classification of barotropic tidal inlets.
Continental Shelf Research 22, 2615–2631.
Huzzey, L.M., Brubaker, J.M., 1988. The formation of long-
itudinal fronts in a coastal estuary. Journal of Geophysical
Research 93, 1329–1334.
Lacy, J.R., Stacey, M.T., Burau, J.R., Monosmith, S.G., 2003.
Interaction of lateral baroclinic forcing and turbulence in an
estuary. Journal of Geophysical Research 108 (C3), 3089.
Lerczak, J.A., Geyer, W.R., 2004. Modeling the lateral circula-
tion in straight, stratified estuaries. Journal of Geophysical
Research 34, 1410–1428.
Li, C., Valle-Levinson, A., 1999. A two-dimensional analytic tidal
model for a narrow estuary of arbitrary lateral depth
variation: the intratidal motion. Journal of Geophysical
Research 104 (C10), 23525–23543.
Lin, J., Kuo, A.Y., 2001. Secondary turbidity maximum in a
partially mixed microtidal estuary. Estuaries 24 (5), 707–720.
Mied, R.P., Handler, R.A., Donato, T.F., 2002. Regions of
estuarine convergence at high Rossby number: a solution in
estuaries with elliptical cross sections. Journal of Geophysical
Research 107 (C11), 3206.
Nunes Vaz, R.A., Simpson, J.H., 1985. Axial convergences in a
well-mixed estuary. Journal of Geophysical Research 103,
637–649.
Pritchard, D.W., 1967. Observations of circulation in coastal
plain estuaries. In: Lauff, G.H. (Ed.), Estuaries. American
Association for the Advancement of Science, pp. 37–44.
Ridd, P.V., Stieglitz, T., Larcombe, P., 1998. Density-driven
secondary circulation in a troptical mangrove estuary.
Estuarine and Coastal Marine Science 47, 621–632.
Sanford, L.P., et al., 2005. Variability of suspended particle
concentrations, sizes and settling velocities in the Chesapeake
Bay turbidity maximum. In: Droppo, I.G., Leppard, G.G.,
Liss, P., Milligan, T. (Eds.), Flocculation in Natural and
Engineered Environmental Systems. CRC Press, Boca Raton,
FL, pp. 211–236.
Simpson, J.H., Brown, J., Mathews, J., Allen, G., 1990. Tidal
straining, density currents, and stirring in the control of
estuarine stratification. Estuaries 13 (2), 125–132.
Swift, M.R., Fredriksson, D.W., Celikkol, B., 1996. Structure of
an axial convergence zone from acoustic doppler current
profiler measurements. Estuarine and Coastal Marine Science
43 (109–122).
Valle-Levinson, A., Boicourt, W.C., Roman, M.R., 2003a. On
the linkages among density, flow, and bathymetry gradients at
the entrance to the Chesapeake Bay. Estuaries 26 (6),
1437–1449.
Valle-Levinson, A., Reyes, C., Sanay, R., 2003b. Effects of
bathymetry, friction, and rotation on estuary–ocean ex-
change. Journal of Physical Oceanography 33 (2375).
Wong, K.-C., 1994. On the nature of transverse variability in a
coastal plain estuary. Journal of Geophysical Research 99
(C7), 14,209–14,222.
ARTICLE IN PRESS
D.C. Fugate et al. / Continental Shelf Research 27 (2007) 679–698698
