Characteristics of a Shallow River Plume: Observations from the Saco River Coastal Observing System

Abstract

Interest in the coastal dynamics of river plumes has mainly focused on large rivers, but plumes from the more numerous smaller
rivers have important local consequences and may, in aggregate, be significant contributors to coastal circulation. We studied
the dynamics of the plume from the Saco River in Saco Bay, Gulf of Maine, over a 3-year period. The transport and salinity
in the region are governed by river discharge, tides, winds, and interaction with the Western Maine Coastal Current. The dynamics
of the flow field vary with location within the plume and discharge. The far-field dynamics of the Saco River plume are dominated
by inertial processes (hence qualifying it as a small-scale river plume), during times of low discharge, with low salinity
water present both up and downstream of the river mouth, but are dominated by rotational processes during times of high discharge
(thus qualifying it as a large-scale river plume), with buoyant water primarily advected downshelf. Near-field dynamics are
governed by weak, subcritical flow during low discharge but strongly inertial, supercritical flow during high discharge. Offshore
movement of the plume is not governed by Ekman dynamics but is instead a result of discharge, wind-induced vertical mixing,
and the geography of the coastline and adjacent islands.

KeywordsRiver plumes–Winds–Tides–Discharge–Observations–Estuaries–Shelf dynamics

Full text

Page 1

Characteristics of a Shallow River Plume: Observations
from the Saco River Coastal Observing System
Charles E. Tilburg & Shaun M. Gill & Stephan I. Zeeman & Amy E. Carlson &
Timothy W. Arienti & Jessica A. Eickhorst & Philip O. Yund
Received: 16 March 2010 /Revised: 3 December 2010 /Accepted: 2 April 2011 /Published online: 3 May 2011
# Coastal and Estuarine Research Federation 2011
Abstract Interest in the coastal dynamics of river plumes
has mainly focused on large rivers, but plumes from the
more numerous smaller rivers have important local con-
sequences and may, in aggregate, be significant contributors
to coastal circulation. We studied the dynamics of the
plume from the Saco River in Saco Bay, Gulf of Maine,
over a 3-year period. The transport and salinity in the
region are governed by river discharge, tides, winds, and
interaction with the Western Maine Coastal Current. The
dynamics of the flow field vary with location within the
plume and discharge. The far-field dynamics of the Saco
River plume are dominated by inertial processes (hence
qualifying it as a small-scale river plume), during times of
low discharge, with low salinity water present both up and
downstream of the river mouth, but are dominated by
rotational processes during times of high discharge (thus
qualifying it as a large-scale river plume), with buoyant
water primarily advected downshelf. Near-field dynamics
are governed by weak, subcritical flow during low
discharge but strongly inertial, supercritical flow during
high discharge. Offshore movement of the plume is not
governed by Ekman dynamics but is instead a result of
discharge, wind-induced vertical mixing, and the geography
of the coastline and adjacent islands.
Keywords River plumes.Winds.Tides.Discharge.
Observations.Estuaries.Shelf dynamics
Introduction
The dynamics of the coastal ocean are inextricably linked to
the terrestrial environment by rivers. River discharges
produce buoyant plumes of freshwater that affect coastal
circulation patterns and may carry nutrients, sediment, and
contaminants that affect coastal food webs (Rabalais et al.
2000). Buoyant river plumes often extend over expansive
areas (Lentz and Limeburner 1995) and affect transport and
mixing for large distances downshelf of the mouth
(Munchow and Garvine 1993). Biological processes in the
coastal ocean are closely coupled to the dynamics of
freshwater plumes. So, an understanding of the effects of
river plumes on the coastal environment requires knowl-
edge of both physical circulation and mixing processes.
The dynamics of river plumes have important implica-
tions for the transport of larvae, contaminants, nutrients,
and pathogens (e.g., Venkatesan et al. 1998; Lipp et al.
2001; Schiff et al. 2003; Gersberg et al. 2004; Warrick et al.
2007; Nagvenkar and Ramalah 2009; Wargo et al. 2009).
Coastal regions can be subject to episodic contamination
during storm events from fecal bacteria emanating from
rivers (Schiff et al. 2003). Understanding the transport of
fecal bacteria within the plume, especially under conditions
in which a shallow plume is advected onto shore, is thus
critical for predicting human impacts along the shore.
Rivers can influence the delivery of nutrients, the timing of
delivery and their relative ratios to coastal waters and thus
affect the productivity and species composition of coastal
phytoplankton (e.g. Otero and Siegel 2004; Warrick et al.
2005). Similarly, toxic algal blooms can occur where
nutrient-rich river plumes contact nearshore cyst beds
(Anderson et al. 2005). Because rainfall and snow melt
are strongly seasonal and exhibit interannual variation, river
plumes and the biological processes that they influence
C. E. Tilburg (*):S. M. Gill:S. I. Zeeman:A. E. Carlson:
T. W. Arienti:J. A. Eickhorst:P. O. Yund
Department of Marine Sciences and Marine Science Center,
University of New England,
11 Hills Beach Road,
Biddeford, ME 04093, USA
e-mail: ctilburg@une.edu
Estuaries and Coasts (2011) 34:785–799
DOI 10.1007/s12237-011-9401-y

Page 2

vary on time scales of days, months, years, and decades
(e.g., Yanagi and Hino 2005; Thomas and Weatherbee
2006). Consequently, the understanding of the fate of
freshwater discharge to the coastal ocean and the causes
and consequences of the freshwater plumes has become a
major goal of coastal oceanography (Henrichs et al. 2000).
While much work has focused on larger rivers (e.g.,
Munchow and Garvine 1993; Rabalais et al. 2000), the
dynamics and characteristics of smaller rivers are still not
well understood (Gaston et al. 2006). However, the
dynamics of smaller river plumes are vital to processes on
the local scales that are relevant to many management
practices and the cumulative effect of the inflows from the
more numerous smaller rivers may be as important as that
of the less numerous larger rivers. Consequently, recent
studies have begun to examine the fate of small rivers as
they enter the coastal region (e.g. Warrick et al. 2007).
This study examines the dynamics of the Saco River
plume, located in Saco Bay in the northwestern Atlantic, a
region governed by wind-driven transport, tidal currents,
and large-scale buoyancy forcing from the Gulf of Maine
coastal current system (Pettigrew et al. 1998, 2005). Here,
we determine the governing characteristics of the flow field
and the physical mechanisms responsible for (1) the across-
shelf extent of the Saco River plume, (2) transport and
salinity within the plume, and (3) vertical mixing of the
plume using data from the Saco River Coastal Observing
System (SaRCOS) sensors and hydrographic transects of
the Saco River plume and the adjacent coastal region.
Background
A number of modeling and theoretical studies have
classified the dynamics responsible for the transport and
mixing of buoyant plumes based on readily measured
criteria (e.g. Garvine 1995; Yankovsky and Chapman 1997;
Avicola and Huq 2001). Garvine (1995) classified buoyant
plumes as “small-scale” or “large-scale” based on the
relative contributions of inertial and rotational processes
within the plume. Large-scale plumes tend to be more
affected by the Earth's rotation than inertial dynamics.
Large-scale plumes leave their source and turn downshelf
(i.e. in the direction of Kelvin wave propagation), creating
a geostrophically balanced coastal current (e.g. Chapman
and Lentz 1994) that transports buoyant water downshelf
of the mouth. In the absence of winds, little buoyant water
is typically found upshelf of the mouth (Munchow and
Chant 2000). Small-scale plumes are governed by inertial
dynamics and tend to form freshwater bulges that radiate in
all directions from the source (Garvine 1995), thus
regularly producing low salinity upshelf of the source.
Garvine (1995) used the bulk Kelvin number, K, a measure
of the relative importance of rotational and inertial
processes to classify the scale of the plumes. The bulk
Kelvin number is expressed as:
K ¼Rp
RD
ð1Þ
where Rpis the observed across-shelf extent of the buoyant
plume or coastal current and RDis the radius of deformation:
RD¼
ffiffiffiffiffiffiffiffi ffi
g0hp
f
p
ð2Þ
where hpis the thickness of the plume, f is the Coriolis
parameter, and g0¼ gðramb? roÞ=rowhere g is gravitation-
al acceleration, ρambis the ambient ocean density, and ρois
the density of the incoming freshwater. Large Kelvin
numbers (K>>1) are indicative of “large-scale” plumes,
and small Kelvin numbers (K<<1) indicate “small-scale”
plumes. While the Kelvin number is largely a geometric
parameter that examines length scales of the plume, the
Rossby number, R, is a dynamic parameter commonly used
to compare inertial and rotational processes of a flow field.
The Rossby number is expressed as:
R ¼
uc
fRD
ð3Þ
where uc is a characteristic velocity of the flow. Large
Rossby numbers (R>>1) indicate flow that is dominated by
inertial processes, while small Rossby numbers (R<<1)
indicate flow dominated by rotational processes.
Since plumes can be affected by winds and other
frictional processes, the Ekman number, Ek, can be used
to estimate the relative strengths of frictional and rotational
processes and can be expressed as:
Ek¼Kz
fh2
p
ð4Þ
where Kzis the vertical eddy viscosity. Large values of the
Ekman number (Ek>1) indicate friction-dominated flow,
while small values (Ek<1) indicate rotational flow
(Cushman-Roisin 1994).
Chao (1988) used the Froude number to characterize the
tendency of flow fields to be governed by baroclinicity or
inertial processes. The Froude number, Fr, is expressed as
Fr ¼
ucffiffiffiffiffiffiffiffi ffi
g0hp
p
ð5Þ
Larger values of the Froude number (Fr>>1) correspond
to supercritical flow in which inertial processes predomi-
nate and stratification processes are not important. Smaller
values (Fr≤1) correspond to subcritical flow in which
stratification is important (Cushman-Roisin 1994).
786 Estuaries and Coasts (2011) 34:785–799

Page 3

Since the dynamics that govern the flow vary depending
on location within the plume (e.g., Chao and Boicourt
1986), these parameters can be calculated within the bulge
of the plume, the coastal current emanating from the bulge,
or the mouth itself. R and Fr can be calculated throughout
the plume to examine how dynamics change within the
plume. However, the mouth Kelvin number, Km, requires a
different length scale (here we use the mouth width, W, as
the characteristic length scale). The mouth Kelvin number
is defined here as (e.g. Huq 2009):
Km¼W
RD
ð6Þ
Tides (e.g. Simpson and Souza 1995; Sanders and
Garvine 2001) can also affect the plume shape, thickness
and location. A useful classification of estuaries by
circulation and tidal mixing parameters is the effective
ratio of the cross-sectionally averaged freshwater velocity
to the average tidal current (Hansen and Rattray 1966). This
tidal index, P, is expressed as:
P ¼
QR=h0W
ut
ðÞ
ð7Þ
where QRis the river discharge, hois the average river
mouth depth, W is the river mouth width, and utis the mean
tidal speed. Large values of the tidal index (P>1) indicate a
buoyancy governed flow, while small values of the tidal
index (P<1) indicate tidally-driven flow.
Yankovsky and Chapman (1997) classified plumes as
“surface-advected,” in which the plume is relatively
shallow and the flow is dictated by surface dynamics, or
“bottom-advected” in which the plume extends to the
bottom and its offshore extent and velocity structure is
dictated by bottom Ekman dynamics and thermal wind
balance. Using only discharge parameters, Yankovsky and
Chapman (1997) were able to predict the fate of waters
within the plumes. They showed that the behavior of the
plume can be predicted based on the relationship between
the predicted depth of the plume, hb, expressed as:
ffiffiffiffiffiffiffiffiffiffi
g0
hb¼
2fQR
s
ð8Þ
andthemeandepthoftheregion.Avalueofhbthat is less than
the mean depth of the region indicates a “surface-advected”
plume and a value of hbthat is greater than the mean depth of
the region indicates a “bottom-advected” plume.
The shallow nature of “surface-advected” plumes can
make them susceptible to mixing and advection from winds
(e.g. Fong and Geyer 2001; Whitney and Garvine 2005;
Pinones et al. 2005). A measure of the competing
mechanisms of advection and mixing can be achieved by
dividing the Rossby number by the Ekman number,
producing a plume Reynolds number:
Rep¼R
Ek
¼uchp2
KzRD
ð9Þ
Large values of the plume Reynolds number (Rep>1)
indicate advection processes are dominant, while small values
(Rep<1) indicate frictional or mixing processes dominate. The
susceptibility of the plume to overturning can be estimated
using a measure of the relative strengths of stratification and
vertical mixing: the Richardson number (e.g. Kato and
Phillips 1969; Kantha et al. 1977), Ri, which is expressed as:
Ri¼g0Δz
ð
where Δu is the absolute difference between velocities at two
different depths and Δz is the distance between the two
depths. Values of Riless than 0.25 typically indicate that the
kinetic energy of the flow can overcome the vertical
stratification and mix the plume (Smyth and Moum 2000;
Sanders and Garvine 2001).
Δu
Þ2
ð10Þ
Materials and Methods
Study Site
Our investigation was conducted over a three year time
period from August 2006 to June 2009 in the mouth of the
Saco River and the adjacent Saco Bay (Fig. 1). The Saco
River is located in the southwestern Gulf of Maine and is
the 6th largest river discharging into the coastal waters of
the Gulf of Maine, with a discharge rate that can exceed
600 m3s−1during spring runoff events. It enters the Gulf of
Maine through a relatively narrow mouth (approximately
250 m) flanked by two jetties, resulting in a discrete point
of entry. The Saco River Estuary is a partially mixed
estuary. Its discharge extends into the relatively open Saco
Bay (Kelley et al. 2005) and is the primary source of
freshwater for that bay. The Saco River typically generates
a plume that is 1–2 m deep and highly mobile. Tidal range
is 2.7 m. The offshore edge of the region is affected by the
southwestward flowing Western Maine Coastal Current
(WMCC), a component of the river-driven coastal current
system located along the western edge of the Gulf of Maine
(Pettigrew et al. 1998; Geyer et al. 2004).
Hydrographic Surveys
We performed approximately weekly surveys of the Saco
River plume and the adjacent coastal region using a
Estuaries and Coasts (2011) 34:785–799787

Page 4

manually deployed Seabird SBE 25 Sealogger CTD
(Conductivity-Temperature-Depth instrument) and SBE
19plus V2 SEACAT CTD to document the vertical
structure (and determine mean depth) of the plume and a
SBE 45 MicroTSG thermosalinograph to document the
horizontal surface structure (and determine the eastward
edge) of the plume. Sampling followed an adaptive strategy
designed to capture the offshore edge of the plume. A total
of 49 surveys were performed between April and October
over the 3-year period. The majority of the surveys
consisted of east–west transects that encompassed the
plume and a segment offshore of the plume; however, a
number of more extensive surveys that encompassed both
the northward and eastward extent of the plume were also
completed. The location and structure of the plume can be
affected by tides (e.g. Garvine 1974) and the strong diurnal
winds (Pinones et al. 2005). Consequently, surveys were
conducted in the morning and afternoon to capture effects
of both the land- and sea-breeze and throughout the tidal
cycle to capture effects of flood, ebb, and slack tides.
Because salinity of the plume and the ambient coastal ocean
varied seasonally, the eastward edge of the plume could not
be identified by a fixed salinity value but was instead
determined by the steepest horizontal gradient in surface
salinity along the east–west transects. The size of the plume
was a strong function of discharge, so we expressed
locations within the plume as functions of Rossby radius
and not set distances. Here, “far-field” refers to parts of the
plume that are located at least 1Rdaway from the mouth,
while “near-field” refers to portions of the plume that are
located within 1Rdfrom the mouth.
Moored Instruments
Time series data used in this study were acquired from the
Saco River Coastal Observing System (SaRCOS), which
consists of a shore-mounted mooring located at the mouth
of the river (not used in this study) and an ocean mooring
(hereafter referred to as the SaRCOS mooring) that was
located approximately 1 km east of Wood Island (Fig. 1).
The ocean mooring consisted of an Aanderaa RCM-9LW
current meter that measured velocity, salinity, and temper-
Distance from shore (km)
Depth (m)
23456789 1011
−5
−4
−3
−2
−1
0
Distance from shore (km)
Depth (m)
23456789 1011
−5
−4
−3
−2
−1
0
14
16
18
20
22
24
26
28
30
cd
Fig. 1 Areal extent of the buoyant plume during a high discharge and
b low discharge. Colored contours represent salinity. Black filled
circles indicate the locations of CTD casts and thermosalinograph
measurements used in the generation of the color contours. The blue
lines in the upper panel indicate the paths of the cross-sectional views
shown in the lower panel. Cross-sectional view of the plume during c
high discharge and d low discharge
788Estuaries and Coasts (2011) 34:785–799

Page 5

ature (at 1 m depth), a SBE 16PIM Seacat CTD that
measured salinity, temperature, pressure, and fluorescence
(colored dissolved organic matter, turbidity, chlorophyll; at
3 m depth), two SBE 37-IM MicroCat CTDs that measured
salinity, temperature, and pressure (at 8 and 12 m depths),
and two SBE 37-IM MicroCat CTs that measured salinity
and temperature (at 5 and 22 m depths). In the spring of
2009, a downward looking 600 kHz RDI Workhorse
Acoustic Doppler Current Profiler (ADCP) was installed
(at 2 m depth) on the mooring to improve the vertical
resolution of the velocities. All mooring data were collected
at 20 min intervals. The conductivity sensor on the
Aanderaa RCM-9LW experienced significant biological
fouling due to the settlement of barnacles and hydroids
each spring and summer. Consequently, the time series for
surface salinity was limited to times before or after the
fouling period or when fouling organisms had been
manually removed. Since the irregular coastline (and
associated islands) did not allow for a straight-forward
determination of along- and across-shelf directions, veloci-
ties were converted to eastward and northward components.
Other Data
Wind measurements in this study were obtained from the
National Oceanic and Atmospheric Administration
(NOAA) environmental buoy EB 44007, located 15 km
northeast of the study site. Wind data were collected at
20 min intervals. Surface salinity measurements were also
obtained from GoMOOS mooring “C,” located 24 km
upshelf of the region within the WMCC (Pettigrew et al.
2005). Wind stress was calculated using the formulation by
Large and Pond (1981). Daily river discharge data were
obtained from the US Geological Survey gauging station at
Cornish, ME. The discharge data were corrected for
drainage areas downstream of the gauging station by
dividing the total drainage area of the watershed by the
discharge area upstream of the gauging station (e.g.
Anderson et al. 2005).
Data Analysis
Salinities, velocities, and winds were examined using
spectral and coherence analysis. The confidence levels
and intervals for correlations, coherence, and spectra
were computed following the methods described by
Emery and Thomson (2001). Every reported correlation
coefficient is significantly different from zero at the 95%
confidence level. For the moored and wind data series, the
degrees of freedom (Neff) replaced the number of obser-
vations (N) in determining significance. Neffwas calculat-
ed by dividing the total time of the observations by the
time of the first zero crossing of the autocorrelation
function (e.g. Bretherton et al. 1999). Principal component
and harmonic tidal analysis were performed on surface
velocity data to determine the direction of maximum
variance and the tidal velocities. To remove high-frequency
diel and tidal variations prior to correlation calculations, the
winds, currents, and salinity data sets were filtered using a
Lanczos low-pass filter with a cut-off frequency of 1/36 h-1
(Jones and Epifanio 1995). Since the region was affected by
large variation in discharge, calculations were performed
for river flows grouped into high discharge (>500 m3/s),
low discharge (<65 m3/s) and moderate discharge (all other
values) periods.
Results
Representative areal plots and vertical cross-sections of the
region show a plume that varies due to discharge (Fig. 1)
and wind direction (Fig. 2). Salinities in the plume during
low discharge varied from a low of 26 at the surface of the
plume to a high of 33 in the ambient waters (right panel of
Fig. 1). Under high discharge conditions, there was much
greater range (both horizontal and vertical) in salinity,
varying between 13 and 33 (left panel of Fig. 1); however,
the plume thickness remained between 1 m and 2 m for all
conditions. The effect of winds on the plume was not
consistent with Ekman dynamics. Surveys of the plume
during periods of comparable discharge and wind speeds,
but different wind directions (see illustrative examples in
Fig. 2), reveal large differences between the offshore extent
of the plume. Northward winds (which would tend to move
the plume northward and offshore in Ekman dynamics)
reduced the offshore extent of the plume (Fig. 2b).
Southward winds (which would tend to move the plume
southward and onshore in Ekman dynamics) increased the
offshore extent of the plume (Fig. 2a).
Examination of the discharge measured at the USGS
gauging station at Cornish, ME (Fig. 3a) reveals strong
temporal variability in river outflow. Discharge varied
between 40 m3/s and 620 m3/s during the study period.
Discharge was greatest during the spring freshet that occurs
during April and May of each year. Additional, smaller
peaks also occurred in the falls of 2006 and 2008 due to
increased storm activity (mainly the remnants of tropical
storms) during these periods. There was also considerable
interannual variability. Conditions were wetter and river
discharge greater in 2006 and 2008 than in 2007.
Although the biological fouling of the Aanderaa con-
ductivity sensor prevented continuous coverage of the
surface salinity, the time series of salinity data at the
different depths did encompass portions of spring, summer,
and winter, and exhibited strong interannual, seasonal, and
daily variability. Salinity trends at three representative
Estuaries and Coasts (2011) 34:785–799 789

Page 6

depths (1, 5, and 22 m depth) are shown in Fig. 3b.
Although additional salinity measurements (at 3, 8, and
12 m) are available, they closely track the representative
depths shown, with salinity at 3 m and 8 m reflecting
salinity at 5 m and salinity at 12 m closely following
salinity at 22 m. This pattern indicates a shallow, surface
plume and little vertical stratification below the plume.
Salinity values were at their lowest during the spring freshet
of each year and surface salinity periodically dropped
below 20. Salinity was significantly higher during the
summers and winters when discharge was at its lowest
(Fig. 3b). The time series of salinity values at the mooring
was consistent with the vertical cross sections, indicating
fresher water at the surface and more saline water at
depth. Comparison of mean salinity during high (dis-
charge >500 m3/s), low (discharge <65 m3/s), and
moderate (all others) river discharge reveals an inverse
relationship between discharge and salinity in the region
Distance from shore (km)
Depth (m)
23456789 1011
−5
−4
−3
−2
−1
0
Distance from shore (km)
Depth (m)
23456789 10 11
−5
−4
−3
−2
−1
0
14
16
18
20
22
24
26
28
30
ab
Fig. 2 Cross-sectional view of the plume during a southward winds and b northward winds
0
200
400
600
800
Discharge (m3/s)
10
15
20
25
30
Salinity (PSU)


Oct Jan Apr Jul Oct Jan Apr Jul Oct Jan Apr Jul
0
2
4
6
8
10
12
Offshore distance
of plume edge (km)
2006200720082009
1 m
5 m
22 m
a
b
c
Fig. 3 Time series of a dis-
charge (m3/s) obtained from the
Cornish, ME USGS gauging
station, b the salinity at depths
of 1, 5, and 22 m at the SaRCOS
mooring, and c offshore location
of the eastern plume boundary
(km) as a function of time. The
lack of discharge data during
late spring 2009 was due to the
presence of ice at the gauging
station
790 Estuaries and Coasts (2011) 34:785–799

Page 7

(Table 1). Low discharge resulted in high salinity and little
vertical stratification, while high discharge resulted in low
salinity and large vertical stratification. Correlations be-
tween salinity and discharge were negative (high discharge
leads to low salinity) as expected; however, the low values
of correlation indicate that other mechanisms affect salinity
fluctuations in the region.
The location of the offshore edge of the plume varied
between the mouth of the river during lowest discharge to
over 10 km offshore during some wind conditions or
extremely high discharge (Fig. 3c). The correlation (with a
1 day lag) between offshore extent of the plume and
discharge was 0.42. To determine the relationship between
winds and plume movement, we compared the offshore
extent or edge of the plume with the projection of the
observed wind velocity vector in each direction. The
highest correlation between wind velocities and offshore
extent (0.45, lag=22 h) occurred when winds were oriented
towards 215°T, which is approximately downshelf.
Examination of the surface velocities revealed a south-
eastward mean flow that varied from 0.08 m s−1during low
discharge to 0.17 m s−1during high discharge (Fig. 4a).
Principal component analysis of velocity revealed that the
greatest variation coincided with the direction of mean flow
and correlated with discharge (Fig. 4b). The proximity of
the mooring to the mouth of the river resulted in mean flow
that was not oriented parallel to the coastline but instead
directed away from the mouth. Harmonic tidal analysis
revealed M2 tidal velocities that were comparable or less
than the mean flow. The major tidal axis (Fig. 4c) did not
coincide with the direction of greatest variation in velocity,
indicating other factors such as discharge and wind-driven
processes affected flow in the region.
Power spectra of salinity time series (Fig. 5a) reveal strong
variation at low frequencies at all depths and at M2 (1/
12.42 h−1) and M4 (1/6.21 h−1) tidal frequencies
(corresponding to 1.9 and 3.9 cpd, respectively) at 1 m and
22 m depths. Interestingly, there was no variation in salinity
at the 5 m depth at the expected tidal frequencies. The
velocity power spectrum (Fig. 5b) shows strong fluctuations
at the tidal frequencies as well as some significant variation
at 0.2–0.3 cpd (3–5 days) that are consistent with wind
fluctuations for the 1 m velocities; however, the velocity
spectra at 5 and 22 m show fluctuations only at tidal
frequencies. The wind velocity power spectrum (Fig. 5c)
shows significant variation only at a diurnal time period
(1 cpd) most likely associated with the sea breeze.
We examined the eastward and northward winds
separately to determine any change in the ocean's response
to the different forcings (Geyer et al. 2004). The winds and
currents were filtered to remove diel and tidal frequency
variations before correlation coefficients were calculated.
Correlations for both components were higher for wind
velocities than wind stress, which is consistent with
previous studies (e.g., Garvine 1991; Whitney and Garvine
2005; Warrick et al. 2007). Eastward wind velocity was
most highly correlated with surface eastward currents (r=
0.62, lag=0 h), consistent with downwind frictional
movement in a nearshore environment (Tilburg 2003).
Northward wind velocity was most highly correlated with
surface currents oriented toward 35°N (r=0.71, lag=0 h),
suggesting a combination of Ekman dynamics and induced
geostrophic motion (Geyer et al. 2004). Linear regression
indicated that the correlated currents were 1.3% and 2.1%
of the eastward and northward wind velocity, respectively.
These relationships are compatible with the “mariner's rule”
that wind-driven currents are a few percent of wind speed
(Tilburg and Garvine 2003). Currents at 4 m and below
were not significantly correlated with winds, indicating that
the depth of penetration of wind stress (<4 m) was
comparable to the plume thickness (1–2 m).
Coherence between salinity variation at 1 m and the
surface currents was significant at the M2 and M4 tidal
periods as well as 3–5 day time scales (0.2–0.3 cpd) of
wind forcing, and longer time scales (<0.05 cpd) most
likely associated with seasonal variation in discharge
(Fig. 6a). Phase spectra were also calculated for coherence
between salinity and surface currents (not shown). Phases
of significant coherence between salinity and velocity
occurred between −140° and −160° for eastward velocities
(maximum coherence between eastward flow and reduced
salinities would result in a phase of −180°) and between
50° and 70° for northward velocities (maximum coherence
between northward velocities and reduced salinities would
result in a phase of 180°), indicating that eastward and
southward (to a lesser extent) velocities (i.e. those that
would transport fresh river water to the mooring) resulted in
decreased salinity. Interestingly, there was no coherence
Table 1 Salinity values at SaRCOS mooring
Depth Low discharge Moderate dischargeHigh discharge Correlation between salinity and dischargeTime lag (days)
1 m
5 m
22 m
30.95
31.54
32.04
29.96
31.31
32.04
27.75
30.38
31.74
−0.33
−0.22
−0.13
0
0
0
Estuaries and Coasts (2011) 34:785–799791

Page 8

−10010
−25
−20
−15
−10
−5
0
5
10
15
20
25
Mean Velocity (cm/s)


High Discharge
Total Discharge
Low Discharge
−2000200
−500
−400
−300
−200
−100
0
100
200
300
400
500
Principal Component Axes
−100 10
−25
−20
−15
−10
−5
0
5
10
15
20
25
Tidal Axes (cm/s)
abc
Fig. 4 Mean velocity (cm/s)
(a), principal component axes
(b), and tidal axes (cm/s) (c) of
the surface velocities at the
SaRCOS mooring
0123456
10
−4
10
−2
10
0
10
2
95% Confidence Interval
Power Spectral Density
(PSU2/cpd)
Salinity Power Spectrum


1m
5 m
22 m
0123456
10−1
100
10
1
10
2
10
3
Power Spectral Density
((cm/s)2/cpd)
Velocity Power Spectrum


95% Confidence Interval
Eastward
Northward
0123456
10−2
10−1
10
0
10
1
102
Cycles per day
Power Spectral Density
((m/s)2/cpd)
Wind Velocity Power Spectrum


95% Confidence Interval
Eastward
Northward
a
b
c
Fig. 5 Power spectra of salin-
ities (a) and velocities (b) at
depths of 1, 5, and 22 m, and
surface wind velocities (c).
Spectra were created using a
Welch spectral estimation meth-
od. Spectral calculations were
made with 95% confidence
intervals
792Estuaries and Coasts (2011) 34:785–799

Page 9

between salinity and surface velocities at 1 cpd that would
be consistent with the effects of the sea breeze (e.g. Pinones
et al. 2005). Although there is energy within the wind field
at 1 cpd (Fig. 5c), the absence of coherence suggests that
the complex geometry of the coastline (Fig. 1a) tends to
complicate the small-scale on- and offshore downwind
movement of the plume, preventing wind-forced diurnal
variation of the salinity within the plume.
The effect of winds on salinity did not extend to greater
depths. There was no coherence between salinity and
velocity at the 3–5 day time scales for depths ≥5 m.
(Fig. 6b and c). Coherence was significant at tidal and
seasonal frequencies for 22 m (Fig. 6c) but only seasonal
frequencies for 5 m (Fig. 6b). Coherence was also
significant at approximately 1.4–1.5 cpd for 5 m, although
it is not clear what forcing drives the high coherence at
these higher frequencies. The coherence between surface
salinity and winds was consistent with wind-driven flow;
the highest (although not significant) coherence occurred at
3–5 day time scales.
Correlations between the salinity at the SaRCOS
mooring and surface salinity at the GoMOOS mooring
“C” (which is located within the WMCC, 24 km upshelf of
the SaRCOS mooring) were significant (for at least some
time lags) at all depths (Fig. 7, Table 2). However the
strongest correlation between mooring “C” and the SaR
COS mooring occurred at the 5 m depth and was higher
than the correlation between any two depths within the
SaRCOS mooring (Table 2). Maximum correlations of
salinity at depths of 1 and 5 m exhibited a time lag of
1.2 days, while correlation of salinity at a depth of 22 m
was highest from 0 to 0.7 days (Fig. 7). The distance
between mooring “C” and the SaRCOS mooring (≈24 km)
should result in sizable time lags in the salinity signal.
Geyer et al. (2004) found that downshelf velocities within
the WMCC varied between 0.2 m/s and 0.4 m/s. Using
0123456
0
0.2
0.4
0.6
0.8
1
Coherence
Coherence of velocities with salinity at 1 m


Eastward
Northward
0123456
0
0.2
0.4
0.6
0.8
1
Coherence
Coherence of velocities with salinity at 5 m
0123456
0
0.2
0.4
0.6
0.8
1
Coherence
Cycles per day
Coherence of velocities with salinity at 22 m
a
b
c
Fig. 6 Coherence of velocities
with salinities at depths of 1 m
(a), 5 m (b), and 22 m (c). Red
lines indicate 95% (solid line)
and 90% (dashed line) signifi-
cance. Levels of significance
were calculated using an
estimated Neff=16
Estuaries and Coasts (2011) 34:785–799793

Page 10

these speeds as estimates of the propagation speed of the
salinity signal in the WMCC, the expected lag time of the
maximum salinity correlation was between 0.7 and 1.4 days,
which encompasses the observed time lags (Fig. 7).
Values of several parameters that describe the general
characteristics of the plume (Garvine 1995; Yankovsky and
Chapman 1997) for high, moderate and low discharge
(Table 3) suggest that changes in river discharge can result
in dramatically different dynamics that govern the propa-
gation and structure of the plume. The ranges in river
discharge result in large variation in the tidal index, P,
indicating that the flow is buoyancy dominated (P>1)
during high discharge and tidally driven (P<1) during low
discharge (Table 3). Examination of K indicates that ranges
in river discharge result in dramatically different dynamics
in the far-field. During high discharge, the plume extends
further offshore (Rp~9 km) and affects a larger portion of
the bay. The greater spatial extent of the plume allows for
rotational processes to dominate (K>1) within the plume
away from the mouth. During low and moderate discharge,
the plume is smaller (Rp<2 km), resulting in inertial
processes dominating throughout the plume (Table 3).
Values of Rossby number are consistent with observed
values of K. For low discharge, R~1, indicating both
inertial and rotational processes were important, but at high
discharge, R<1, indicating rotational processes dominated
within the plume. Using a typical value of vertical eddy
viscosity (~2.0×10−4m2/s) found within coastal river
plumes (e.g. Houghton et al. 2004; Fong and Geyer
2001), the Ekman number was calculated to be 0.5,
indicating that rotational processes predominate over
frictional processes. Examination of Fr reveals that strati-
fication was important at all times within the plume.
Dynamics near the mouth differed from those within the
body of the plume. The small inlet of the Saco River (W=
250 m) resulted in small values of Kmfor all discharges
(Table 3, 0.075–0.17), indicating that the region near the
mouth is affected by inertial processes and the discharge is
likely to form a bulge (Geyer et al. 2004; Garvine 1999)
with low salinities both up- and downstream of the mouth.
However, values of Rm were small, indicating that
rotational processes were also important near the mouth.
Examination of Frm reveals that during high discharge,
inertial processes were more important than stratification
near the mouth, but at lower discharges, stratification was
more important in governing the flow.
Calculated values of hb(Yankovsky and Chapman 1997)
are consistent with the observed thickness of the plume and
indicate a “surface-advected” plume for all discharge values
(Table 3). The shallow nature of the plume (~ 1–2 m depth)
can make it susceptible to mixing and advection from
winds. Plume Reynolds numbers were near 1 for all values
of discharge indicating that both advection and mixing
processes can be important. Richardson numbers were
calculated using the density and velocity at 1 m and 4 m
depths. Time series of the projection of wind velocity onto
a vector oriented towards 215°T (Fig. 8a) and Richardson
number (Fig. 8b) during an example period of moderate
012345678910
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Lag (days)
Correlation
Lagged correlation of salinities with GoMOOS Mooring "C" salinity


1 m
5 m
22 m
Fig. 7 Lagged correlation be-
tween salinity at different depths
at the SaRCOS mooring and
surface salinity at GoMOOS
mooring “C.” Horizontal red
dashed line indicates 95% sig-
nificance. Black vertical line
indicates time lag of maximum
correlation for 1 and 5 m salin-
ities. Black vertical dashed lines
represent minimum and
maximum expected time lag
calculated from propagation
speeds in the WMCC
794Estuaries and Coasts (2011) 34:785–799

Page 11

discharge (March 2009) show that during strong wind
events Ri frequently falls below 0.25, indicating mixing of
the plume with the ambient ocean. During high discharge
(not shown), Ri rarely falls below 0.25, indicating that high
discharge events do not typically mix the plume, despite the
high Fr near the mouth.
Discussion
Characteristics of the Buoyant Flow
Observations and calculations indicated a shallow plume
whose salinity and flow are strongly influenced by winds,
tides, discharge, and the presence of the Western Maine
Coastal Current. The size and character of the plume was a
strong function of discharge. The dynamics of the plume
changed from “small-scale,” in which inertial effects
dominated flow patterns throughout the plume during low
discharge, to “large-scale,” where rotational effects domi-
nated in the far-field (i.e. >1Rdaway from mouth) during
high discharge. Due to the small inlet, the effects of
discharge on the dynamics of the flow were different near
the mouth, where high discharge produced strongly inertial,
supercritical flows, and low discharge resulted in subcritical
flows. Low discharge resulted in a tidally-driven flow near
the mouth, while high discharge resulted in a buoyancy-
driven flow. Calculations of hb revealed that the plume
would be “surface-advected” for all discharge regimes
(Table 3). The plume was rarely thicker than 1–2 m deep
(indicating a “surface-advected” plume). Currents in the
region were highly correlated with winds, revealing both a
simple downwind (for eastward) and Ekman (for northward)
response to the winds; however, the effects of winds were
constrained to the surface due to the strong vertical
stratification within the plume. The areal and vertical
structure of the plume (Fig. 1) is consistent with the predicted
characteristics of the plume. Low salinities were frequently
detected up- and downshelf of the mouth (indicating near-
field bulge formation); however, during high discharge, low
salinities were found slightly upshelf but more than 4 km
downshelf (indicating far-field geostrophic turning).
Interaction with the Western Maine Coastal Current
The significant coherence of surface velocity and salinity at
the mooring at tidal frequencies indicates that the salinity is
highly affected by tidal flow. Interestingly, the high
coherence is confined to the shallowest (1 m) and deepest
mooring (22 m) depths. Although the middle depth (5 m)
exhibits strong variation in velocity at tidal frequencies
(Fig. 5b), there is little variation in salinity (Fig. 5a) and no
coherence at tidal frequencies (Fig. 6). The barotropic
nature of tides should result in high correlations between
salinity and tidal velocity at all depths in a region heavily
Table 2 Correlation values of salinities at SaRCOS and GoMOOS
“C” moorings
1 m5 m22 m
“C”
1 m
5 m
22 m
“C”
10.63
1
0.39
0.71
1
0.51
0.79
0.61
1
“C” mooring salinity was measured at a depth of 1 m
Parameter SymbolLow dischargeModerate discharge High discharge
Discharge (m3/s)
Tidal index
Reduced gravity (N/kg)
Internal Rossby Radius (km)
Observed radius (km)
Mean speed (m/s)
Bulk Kelvin number
Mouth Kelvin number
Rossby number
Mouth Rossby number
Ekman number
Froude number
Mouth Froude number
Plume Reynolds number
Predicted depth (m)
Qr
P
g′
RD
RP
uave
K
Km
R
Rm
Ek
Fr
Frm
Rep
hb
65110500
0.56
0.015
1.5
0.8
0.082
0.53
0.17
1.02
0.62
0.50
0.54
0.33
1.09
0.81
1.2
0.036
2.3
1.6
0.11
0.69
0.11
0.67
0.68
0.50
0.46
0.47
0.95
0.78
4.4
0.075
3.4
9.0
0.18
2.7
0.075
0.19
0.44
0.50
0.52
1.19
1.06
1.0
Table 3 Key parameters of the
Saco River Plume
f=9.9×10−5s−1, W=250 m
Estuaries and Coasts (2011) 34:785–799795

Page 12

influenced by river discharge. However, examination of the
large-scale flow field reveals that the offshore edge of the
region is dominated by the southwestward flowing,
baroclinic WMCC (Geyer et al. 2004; Churchill et al.
2005) whose flow is governed by geostrophic and wind-
driven transport (Geyer et al. 2004; Hetland and Signell
2005). Recent studies of the WMCC show that it can
penetrate shoreward of the SaRCOS mooring and generally
extends to depths of 15–25 m (Geyer et al. 2004; Pettigrew
et al. 2005). Salinity correlations between the SaRCOS
mooring and the upshelf mooring “C” were stronger at 5 m
(r=0.79) than at 22 m depth (r=0.61), suggesting that the
SaRCOS mooring is strongly affected by the baroclinic
WMCC, which does not always extend to the bottom in this
region. The lack of coherence of salinity with tides at 5 m is
consistent with the results of Geyer et al. (2004) who found
weak tidal effects in the WMCC. Coherence of salinity with
tides at 22 m (but not at 5 m) suggests that the WMCC that
does not always extend to 22 m. The lack of coherence of
salinity with winds at 5 and 22 m is consistent with the
shallow Ekman depths created by the buoyant discharge
from the Saco River. In contrast to our study, Geyer et al.
(2004) found that the WMCC is strongly influenced by
winds; however, their analysis did not extend to the highly
stratified inshore regions such as those affected directly by
the Saco River discharge.
Offshore Movement of the Plume
The offshore movement of the eastward edge of the
buoyant plume is governed by the salinity field created by
river discharge and vertical mixing due to winds. Higher
discharge results in greater offshore movement, which is
consistent with a small Km and larger Frm, resulting in
buoyant transport from the mouth that leads to bulge
formation. Garvine (1999) showed that for small Km, the
discharge should create a bulge whose horizontal scale
varies with river discharge. The positive, lagged correlation
between discharge and offshore movement of the plume is
consistent with the creation and subsequent expansion of a
buoyant bulge of freshwater emanating from the mouth
(Nof and Pichevin 2001).
The offshore movement of the plume due to winds is not
a product of Ekman dynamics or even downwind advection
of the plume. Plume Reynolds numbers of approximately
unity for all discharges suggests that mixing due to winds is
as important as advection. The highest correlation between
offshore movement of the plume and winds (r=0.45, lag=
22 h) occurs when the wind component is oriented towards
215°T, which is not consistent with Ekman dynamics (in
which maximum correlation would occur for northeastward
winds) or downwind frictional advection (in which maxi-
mum correlation would occur for eastward winds). The
relationship between winds and the offshore movement of
the plume appears to be based on physical mechanisms that
take time to develop; the time lag between winds and
offshore plume extent (22 h) is greater than that expected
for Ekman dynamics (Garvine 1991) in a region where the
highest correlation between velocities and winds shows no
time lag. There is not a significant correlation between the
velocities at the mooring and the offshore extent of the
plume, which suggests a physical mechanism other than
1113 1517192123 25 2729 31
−10
−5
0
5
10
Wind speed (m s−1)
1113 1517 192123 25272931
0
0.2
0.4
0.6
0.8
1
Richardson Number
March, 2009
a
b
Fig. 8 a Projection of wind
velocity onto a vector oriented
toward 215°T. Positive indicates
southwestward flow. Negative
indicates northeastward flow. b
Richardson number at SaRCOS
mooring. Red dashed line
indicates a Richardson
number of 0.25
796Estuaries and Coasts (2011) 34:785–799

Page 13

advection governing the edge of the plume. Since the
plume edge results from a balance between the horizontal
transport of buoyancy and the vertical mixing (Fong and
Geyer 2001; Houghton et al. 2004), wind-driven vertical
mixing can result in the eventual destruction of the plume
edge, decreasing the offshore extent of the plume. Calcu-
lations of the Richardson number (Fig. 8) reveal that
commonly encountered winds can mix the shallow plume
with the ambient ocean. However, the plume extent is not
governed simply by wind-mixing, since high discharge
events that tend to produce large plumes are characterized
by strong stratification that limits the ability of winds to
mix the plume. The orientation of the coastline, the location
of the associated islands, and the direction of the dominant
flow field determine the direction of the winds that are most
likely to mix the plume. A significant correlation between
plume extent and the wind velocity component oriented
towards 215°T shows that northeastward winds are most
effective in vertically mixing the plume (reducing its
across-shelf extent), while southwestward winds are the
least effective in mixing the plume. Northeastward winds
oppose the typical flow of the southwestward WMCC
resulting in greater vertical shear and more mixing.
Southwestward winds blow in the same direction as the
WMCC, resulting in less vertical shear, less vertical mixing,
and more advection of the plume.
The tombolo south of the Saco River mouth and Wood
Island (which extend eastward to a longitude of 70.3°W or
more than 6 km east of the main coast, labeled in Fig. 1)
determine the extent to which the northeastward winds can
mix the plume and modify the movement of the plume by
southwestward winds. Examination of Fig. 2b reveals that
northeastward winds mix the plume east of Wood Island,
while the portion of the plume that resides in the lee of the
island and tombolo is still intact. The presence of these
features effectively shelters the plume from the mixing
action of the northeastward winds. However, the tombolo
and Wood Island play different roles for southwestward
winds. Typically, downwelling winds tend to generate a
strong downshelf flow while trapping a buoyant plume
against the coast. Choi and Wilkin (2007), in a recent
modeling study of the Hudson River plume, found that
downwelling winds would decrease the across-shore extent
of the plume by a factor of 2. Lentz and Largier (2006) in a
study of the Chesapeake Bay plume showed that the across-
shelf extent of the plume ranged from 2 to 3RDduring
quiescent periods or upwelling events to less than 1RD
during downwelling events. For the Saco River plume, RD
varied between 1.5 km and 3.4 km. However, the presence
of Wood Island interrupts this downshelf flow and forces
the southwestward wind-driven flow to extend offshore of
Wood Island before continuing downshelf. Since the
seaward side of Wood Island is more than 6 km offshore
(i.e. ~ 2–4RD) of the river mouth, a plume that extends only
1RDfrom shore is located much further eastward than a
plume during quiescent or upwelling winds. The downshelf
and subsequent offshore movement takes time to develop,
consistent with the long lag time (22 h) between the onset
of winds and offshore movement.
Summary and Conclusions
In this study, we examined the dynamics of the flow field of
the Saco River plume in Saco Bay, Gulf of Maine using a
combination of moored instruments and weekly transects of
the region. Examination of observations of the plume
revealed a shallow plume that is strongly influenced by
winds, tides, discharge and the coastal current system along
the western edge of the Gulf of Maine. The scale of the
plume and therefore, the governing dynamics, were highly
dependent on discharge. During high discharge, the spatial
scales of the plume were larger than the radius of
deformation, the effect of rotation was enhanced, and the
far-field portions of the plume were affected by both inertial
and rotational processes. During low discharge, the spatial
scales of the plume were small, the effect of rotation was
reduced, and the plume was dominated by inertial processes
throughout the plume. Examination of characteristic
parameters showed that the governing dynamics of the
flow differed depending on the location within the plume.
Near the mouth, high discharge resulted in inertial,
supercritical flow, while low discharge resulted in weaker,
subcritical flow.
The location of the plume was highly dependent on the
strong tidal velocities in the region and the presence of the
southwestward flowing WMCC that was present along the
eastern edge of Saco Bay. The shallow plume was strongly
affected by winds, although not by Ekman dynamics. The
location of the plume was a product of both wind-induced
vertical mixing and the presence of the island and tombolo
located offshore of the river mouth. Strong northeastward
winds, which would tend to transport the plume upshelf and
offshore due to Ekman dynamics, instead mixed the plume
with the ambient ocean, effectively reducing the across-
shore extent of the plume. Southwestward winds would
tend to result in downwelling and downshelf flow;
however, the presence of Wood Island and a tombolo to
the south redirected the downshelf flow to the east before it
returned downshelf, effectively extending the eastward
extent of the plume.
This study suffers from a number of limitations that can
affect the interpretation of the measurements. Moored
observations were confined to only one location within
the plume and biological fouling did reduce the times at
which surface data were available. Transects were confined
Estuaries and Coasts (2011) 34:785–799 797

Page 14

to spring, summer, and fall periods and did not necessarily
coincide with mooring observations. Wave-driven transport
was not addressed in this study but could be important to
plume dynamics and dispersal of pathogens and toxins
(Svejkovsky et al. 2010). However, our results present new
insight into those physical mechanisms that can advect and
mix plumes as well cause plumes to vary between “small-
scale” and “large-scale” dynamics.
Acknowledgments
ogy Institute for funding the equipment and the University of New
England for funding the operating costs of the project. We would like
to thank Michael Dunnington for help with the initial planning of the
project and Kelly Provost, Christina Guidoboni, Theresa Robitaille,
and Emily Zimmerman for assisting in the field. We are also grateful
for the comments by two anonymous reviewers. This is contribution
number 34 from the Marine Science Center at the University of New
England.
The authors wish to thank the Maine Technol-
References
Anderson, D.M., B.A. Keafer, W.R. Geyer, R.P. Signell, and T.C.
Loder. 2005. Toxic Alexandrium blooms in the western Gulf of
Maine: The plume advection hypothesis revisited. Limnology and
Oceanography 50: 328–345.
Avicola, G., and P. Huq. 2001. Scaling analysis for the interaction
between a buoyant coastal current and the continental shelf:
Experiments and observations. Journal of Physical Oceanogra-
phy 18: 1144–1166.
Bretherton, C.S., M. Widmann, V.P. Dyminikov, J.M. Wallace, and I.
Blade. 1999. The effective number of spatial degrees of freedom
of a time-varying field. Journal of Climate 12: 1990–2009.
Chao, S.-Y. 1988. Wind-driven motion of estuarine plumes. Journal of
Physical Oceanography 18: 1144–1166.
Chao, S.-Y., and W.C. Boicourt. 1986. Onset of estuarine plumes.
Journal of Physical Oceanography 16: 2137–2149.
Chapman, D.C., and S.J. Lentz. 1994. Trapping of a coastal density
front by the bottom boundary layer. Journal of Physical
Oceanography 24: 1464–1479.
Choi, B.-J., and J.L. Wilkin. 2007. The effect of wind on the dispersal
of the Hudson River plume. Journal of Physical Oceanography
37: 1878–1897.
Churchill, J.H., N.R. Pettigrew, and R.P. Signell. 2005. Structure and
variability of the Western Maine Coastal Current. Deep-Sea
Research Part II 52: 2392–2410.
Cushman-Roisin, B. 1994. Introduction to Geophysical Dynamics.
Englewood Cliffs, NJ: Prentice Hall.
Emery, W.J., and R.E. Thomson. 2001. Data Analysis Methods in
Physical Oceanography, 2nd edition. Amsterdam: Elsevier
Science.
Fong, D.A., and W.R. Geyer. 2001. Response of a river plume during
an upwelling favorable wind event. Journal of Geophysical
Research 106: 1067–1084.
Garvine, R.W. 1974. Physical features of the Connecticut River
outflow during high discharge. Journal of Geophysical Research
79: 831–846.
Garvine, R.W. 1991. Sub-tidal frequency estuary-shelf interaction:
Observations near Delaware Bay. Journal of Geophysical
Research 96: 7049–7064.
Garvine, R.W. 1995. A dynamical system for classifying buoyant
coastal discharges. Continental Shelf Research 15: 1585–1596.
Garvine, R.W. 1999. Penetration of buoyant coastal discharge onto the
continental shelf: A numerical model experiment. Journal of
Physical Oceanography 29: 1892–1909.
Gaston, T.F., T.A. Schlacher, and R.M. Connolly. 2006. Flood
discharges of a small river into open coastal waters: Plume
traits and material fate. Estuarine, Coastal and Shelf Science
69: 4–9.
Gersberg, R.M., D. Daft, and D. Yorkey. 2004. Temporal pattern of
toxicity in runoff from the Tijuana River Watershed. Water
Research 38: 559–568.
Geyer, W.R., R.P. Signell, D.A. Fong, J. Wang, D.M. Anderson, and
B.A. Keafer. 2004. The freshwater transport and dynamics of the
western Maine coastal current. Continental Shelf Research 24:
1339–1357.
Hansen, D.V., and M. Rattray Jr. 1966. New dimensions in estuary
classification. Limnology and Oceanography 11: 319–326.
Henrichs, S., N. Bond, R. Garvine, G. Kineke, and S. Lohrenz. 2000.
Coastal ocena processes CoOP: Transport and transformation
processes over continental shelves with substantial freshwater
inflows. Report on the CoOP Buoyancy-Driver Transport
Processes Workshop.
Hetland, R., and R.P. Signell. 2005. Modeling coastal current transport
in the Gulf of Maine. Deep Sea Research Part II 52: 2430–2449.
Houghton, R.W., C.E. Tilburg, R.W. Garvine, and A. Fong. 2004.
Delaware River plume response to a strong upwelling favorable
wind event. Geophysical Research Letters 31: L07302.
Huq, P. 2009. The role of Kelvin number on bulge formation from
estuarine buoyant outflows. Estuaries and Coasts 32: 709–719.
Jones, M.B., and C.E. Epifanio. 1995. Metamorphosis of brachyuran
megalopae in Delaware Bay: An analysis of time series data.
Marine Ecology Progress Series 125: 67–76.
Kantha, L.H., O.M. Phillips, and R.S. Azad. 1977. On turbulent
entrainment at a stable density interface. Journal of Fluid
Mechanics 79: 753–768.
Kato, H., and O.M. Phillips. 1969. On the penetration of a turbulent
layer into stratified fluid. Journal of Fluid Mechanics 37: 643–
655.
Kelley, J.T., D.C. Barber, D.F. Belknap, D.M. Fitzgerald, S. van
Heteren, and S.M. Dickson. 2005. Sand budgets at geological,
historical and contemporary time scales for a developed beach
system, Saco Bay, Maine, USA. Marine Geology 214: 117–
142.
Large, W.G., and S. Pond. 1981. Open ocean momentum flux
measurements in moderate to strong winds. Journal of Physical
Oceanography 11: 324–336.
Lentz, S.J., and J. Largier. 2006. The influence of wind forcing on the
Chesapeake Bay buoyant coastal current. Journal of Physical
Oceanography 36: 1305–1316.
Lentz, S.J., and R. Limeburner. 1995. The Amazon River plume
during AMASSEDS: Spatial characteristics and salinity varia-
bles. Journal of Geophysical Research 100: 2355–2375.
Lipp, E.K., R. Kurz, R. Vincent, C. Rodriguez-Palacios, S.R. Farrah,
and J.B. Rose. 2001. The effects of seasonal variability and
weather on microbial fecal pollution and enteric pathogens in a
subtropical estuary. Estuaries 24: 266–276.
Munchow, A., and R.J. Chant. 2000. Kinematics of inner shelf
motions during the summer stratified season off New Jersey.
Journal of Physical Oceanography 30: 247–268.
Munchow, A., and R.W. Garvine. 1993. Buoyancy and wind forcing
of a coastal current. Journal of Marine Research 51: 293–322.
Nagvenkar, G.S., and N. Ramalah. 2009. Abundance of sewage-
pollution indicator and human pathogenic bacteria in a tropical
estuarine complex. Environmental Monitoring and Assessment
155: 245–256.
Nof, D., and T. Pichevin. 2001. The ballooning outflows. Journal of
Physical Oceanography 31: 3045–3058.
798Estuaries and Coasts (2011) 34:785–799

Page 15

Otero, M.P., and D.A. Siegel. 2004. Spatial and temporal character-
istics of sediment plumes and phytoplankton blooms in the Santa
Barbara Channel. Deep Sea Research Part II 51: 1129–1149.
Pettigrew, N.R., D.W. Townsend, H. Xue, J.P. Wallings, P.J. Brickley,
and R.D. Hetland. 1998. Observations of the Eastern Maine
Coastal current and its offshore extensions. Journal of Geophys-
ical Research 103: 30623–30639.
Pettigrew, N.R., J.H. Churchill, C.D. Janzen, L. Mangum, R.P.
Signell, A. Thomas, D.W. Townsend, J.P. Wallinga, and H.
Xue. 2005. The kinematic and hydrographic structure of the Gulf
of Maine Coastal Current. Deep Sea Research Part II 52: 2369–
2391.
Pinones, A., A. Valle-Levinson, D.A. Narvaez, C.A. Vargas, S.A.
Navarrete, G. Yuras, and J.C. Castilla. 2005. Wind-induced
diurnal variability in river plume motion. Estuarine, Coastal and
Shelf Science 65: 513–525.
Rabalais, N.N., R.E. Turner, J. Dubravko, W. Dortch, W.J.
Wiseman Jr., and B.K. Sen Gupta. 2000. Gulf of Mexico
biological system responses to nutrient changes in the
Mississippi River. In Estuarine science: A synthetic approach
to research and practice, ed. J.E. Hobbie, 241–268. Wash-
ington, DC: Island.
Sanders, T.M., and R.W. Garvine. 2001. Fresh water delivery to the
continental shelf and subsequent mixing: an observational study.
Journal of Geophysical Research 106: 27087–27101.
Schiff, K.C., J. Morton, and S.B. Weisberg. 2003. Retrospective
evaluation of shoreline water quality along Santa Monica Bay
beaches. Marine Environmental Research 56: 245–253.
Simpson, J.H., and A.J. Souza. 1995. Semi-diurnal switching of
stratification in the Rhine ROFI. Journal of Geophysical
Research 100: 7037–7044.
Smyth, W.D., and J.N. Moum. 2000. Anisotropy of turbulence in
stably stratified mixing layers. Physical Fluids 12: 1343–1362.
Svejkovsky, J., N.P. Nezlin, N.M. Mustain, and J.B. Kuma. 2010.
Tracking stormwater discharge plumes and water quality of the
Tijuana River with multispectral aerial imagery. Estuarine,
Coastal and Shelf Science (in press).
Thomas, A.C., and R.A. Weatherbee. 2006. Satellite-measured
temporal variability of the Columbia River plume. Remote
Sensing and the Environment 100: 167–178.
Tilburg, C.E. 2003. Across-shelf transport on a continental shelf: do
across-shelf winds matter? Journal of Physical Oceanography
33: 2675–2688.
Tilburg, C.E., and R.W. Garvine. 2003. Three-dimensional flow in a
shallow coastal upwelling zone: Alongshore convergence and
divergence on the New Jersey shelf. Journal of Physical
Oceanography 33: 2113–2125.
Venkatesan, M.I., N. Chalaux, J.M. Bayona, and E. Zeng. 1998.
Butyltins in sediments from Santa Monica and San Pedro basins,
California. Environmental Pollution 99: 263–269.
Wargo, A.M., C.E. Tilburg, W.B. Driggers, and J.A. Sulikowski. 2009.
Effect of a freshwater plume on icthyoplankton distribution off the
coast of southern Maine. Northeastern Naturalist 16: 647–654.
Warrick, J.A., L. Washburn, M.A. Brzezinski, and D.A. Siegel. 2005.
Nutrient contributions to the Santa Barbara Channel, California,
from the ephemeral Santa Clara River. Estuarine, Coastal and
Shelf Science 62: 559–574.
Warrick, J.A., P.M. DiGiacomo, S.B. Weisberg, N.P. Nezlin, M.
Mengel, B.H. Jones, J.C. Ohlmann, L. Washburn, E.J. Terrill, and
K.L. Farnsworth. 2007. River plume patterns and dynamics
within the Southern California Bight. Continental Shelf Research
27: 2427–2448.
Whitney, M.M., and R.W. Garvine. 2005. Wind influence on a coastal
buoyant outflow. Journal of Geophysical Research Oceans 110:
C03014.
Yanagi, T., and T. Hino. 2005. Short term, seasonal, and tidal
variations in the Yellow River plume. Mer 43: 1–7.
Yankovsky, A.E., and D.C. Chapman. 1997. A simple theory for the
fate of buoyant coastal discharges. Journal of Physical Ocean-
ography 27: 1386–1401.
Estuaries and Coasts (2011) 34:785–799799