On the Longshelf Structure and Dynamics of Subtidal Currents on the Eastern United States Continental Shelf

Abstract

Strong correlations were observed among subtidal longshelf currents from the Middle Atlantic Bight (MAB) to the Georges Bank region, a distance spanning 615 km. The longshelf current consisted predominantly of wind-forced motions and freely propagating events, which together accounted for 75%–90% of the longshelf current energy. Much stronger longshelf currents were observed in the MAB than on Georges Bank. The MAB/Georges Bank energy ratio for wind-forced currents on the 60 m isobath was 20. The ratio for freely propagating events was 3. The magnitudes of many of the terms in the vertically integrated wind-driven momentum equations were estimated from observations of current, pressure and surface stress, and from calculations of bottom stress. The cross-shelf momentum balance was geostrophic. Surface and bottom stress, the longshelf pressure gradient, and the Coriolis force on the cross-shelf flow were important terms in the longshelf momentum balance. An analytic model of wind-forced current, which incorporates the significant force balances, accounted for the observed longshelf variation of the wind-forced currents. Average bottom-drag and bottom-resistance coefficients estimated from current and bottom-stress records range from 4–8 (× 10⁻³) and 0.07–0.20 cm s⁻¹, respectively.

Full text

Seasonal Variations in the Circulation over the Middle Atlantic Bight Continental Shelf
STEVEN J. LENTZ
Woods Hole Oceanographic Institution, Woods Hole, Massachusetts
(Manuscript received 24 January 2007, in final form 20 November 2007)
ABSTRACT
Fits of an annual harmonic to depth-average along-shelf current time series longer than 200 days from 27
sites over the Middle Atlantic Bight (MAB) continental shelf have amplitudes of a few centimeters per
second. These seasonal variations are forced by seasonal variations in the wind stress and the cross-shelf
density gradient.
The component of wind stress that drives the along-shelf flow over most of the MAB mid- and outer shelf
is oriented northeast–southwest, perpendicular to the major axis of the seasonal variation in the wind stress.
Consequently, there is not a significant seasonal variation in the wind-driven along-shelf flow, except over
the southern MAB shelf and the inner shelf of New England where the wind stress components forcing the
along-shelf flow are north–south and east–west, respectively.
The seasonal variation in the residual along-shelf flow, after removing the wind-driven component, has
an amplitude of a few centimeters per second with maximum southwestward flow in spring onshore of the
60-m isobath and autumn offshore of the 60-m isobath. The spring maximum onshore of the 60-m isobath
is consistent with the maximum river discharges in spring enhancing cross-shelf salinity gradients. The
autumn maximum offshore of the 60-m isobath and a steady phase increase with water depth offshore of
Cape Cod are both consistent with the seasonal variation in the cross-shelf temperature gradient associated
with the development and destruction of a near-bottom pool of cold water over the mid and outer shelf
(“cold pool”) due to seasonal variations in surface heat flux and wind stress.
1. Introduction
There is a mean equatorward along-shelf flow of
5–10 cm s
⫺1
over the continental shelf of the Middle
Atlantic Bight (MAB) and southern flank of Georges
Bank (e.g., Bumpus 1973; Beardsley and Boicourt 1981;
Butman and Beardsley 1987; Lentz 2008). One might
expect a corresponding seasonal variation (annual
cycle) in the along-shelf flow because there are well-
documented seasonal variations in both the wind stress
(Saunders 1977) and in the structure of the density field
(Bigelow 1933; Bigelow and Sears 1935; Beardsley et al.
1985; Flagg 1987; Linder and Gawarkiewicz 1998).
However, the magnitude, pattern, and cause of sea-
sonal variations in the circulation of the MAB remain
unclear.
There is clear evidence of a seasonal variation in the
along-shelf flow over the southern flank of Georges
Bank with an amplitude of a few centimeters per sec-
ond and maximum southwestward along-shelf flow in
September (Butman and Beardsley 1987; Brink et al.
2003; Flagg and Dunn 2003). Shearman and Lentz
(2003) observed a maximum monthly mean along-shelf
flow of about 10 cm s
⫺1
in September 1996, decreasing
to a few centimeters per second in spring 1997 over the
mid and outer New England shelf. However, several
previous studies of mid and outer-shelf currents in the
same region did not observe a well-defined seasonal
variation in the along-shelf flow (Mayer et al. 1979;
Beardsley et al. 1985; Aikman et al. 1988). In an analy-
sis of current profiles from 10 years of weekly cross-
shelf transects between New York and Bermuda, Flagg
et al. (2006) observed a large seasonal variation in the
shelf–slope jet offshore of the 100-m isobath, with
strongest along-slope flow in winter (20 cm s
⫺1
) and
weaker flow in spring–summer (10 cm s
⫺1
). Linder and
Gawarkiewicz (1998) observed a similar seasonal varia-
tion in geostrophic estimates of shelf–slope jet trans-
port south of Cape Cod. However, Flagg et al. (2006)
observed “little seasonal variation” in the along-shelf
flow onshore of the shelf–slope jet (water depths less
Corresponding author address: Steven J. Lentz, Woods Hole
Oceanographic Institution, MS 21, Woods Hole, MA 02543.
E-mail: slentz@whoi.edu
1486 JOURNAL OF PHYSICAL OCEANOGRAPHY VOLUME 38
DOI: 10.1175/2007JPO3767.1
© 2008 American Meteorological Society
JPO3767

than 100 m). Ullman and Codiga (2004) and Codiga
(2005) observed a substantial seasonal variation in the
along-shelf flow over the inner shelf near the mouth of
Long Island Sound. In contrast to the outer-shelf sites
listed above, the along-shelf flow at this inner-shelf site
was strongest in summer and weak, or nonexistent, in
winter.
Seasonal variations in the along-shelf flow over the
MAB shelf have often been attributed to seasonal
variations in the cross-shelf density gradient (Butman
1987; Shearman and Lentz 2003; Flagg et al. 2006). Ull-
man and Codiga (2004) and Codiga (2005) found that
wind forcing also contributed to seasonal variations in
the along-shelf flow over the inner shelf near Long Is-
land Sound. Several sources of seasonal variation in the
cross-shelf density structure have been identified. In
the vicinity of the 50-m isobath over Georges Bank, the
seasonal variation in the cross-shelf density gradient is
associated with the establishment of a tidal mixing front
in spring and summer (Flagg 1987). Over the New
England shelf, Shearman and Lentz (2003) argue that
seasonal variations in the temperature structure associ-
ated with surface heating and cooling (Bigelow 1933;
Beardsley and Boicourt 1981) cause seasonal variation
in the cross-shelf density gradient. Ullman and Codiga
(2004) and Codiga (2005) found that cross-shelf density
gradients over the inner shelf near Long Island Sound
are associated with buoyant outflow from Long Island
Sound.
The goal of this study is to determine whether there
is a consistent pattern to seasonal variations in the
MAB circulation and, if so, what drives the seasonal
variability. To address these questions, an annual cycle
is fit to current time series longer than 200 days from 27
sites over the continental shelf of the MAB and south-
ern flank of Georges Bank. The analysis indicates that
there are significant seasonal variations in the along-
shelf flow associated with seasonal variations in both
the wind stress and the cross-shelf density gradient.
2. Observations and processing
a. Observations
The 27 sites with long current records span the inner,
mid, and outer shelf, but the spatial coverage is sparse
and the along-shelf distribution is not uniform, with
many of the stations concentrated over the New En-
gland shelf (Fig. 1; Table 1). In particular, there are
relatively few sites south of the Hudson shelf valley,
which extends across the shelf. Three sites from the
southern flank of Georges Bank are included because
the shelf is continuous with the mid- and outer MAB
FIG. 1. Map of the Middle Atlantic Bight showing the locations of the current-meter moor-
ing sites (solid circles), wind stations (solid triangles), and seasonal ellipses for depth-averaged
currents (blue) and wind stress (red). Not all ellipses are shown for clarity. The 50-, 100-, and
1000-m isobaths are shown.
JULY 2008 L ENTZ 1487
Fig 1 live 4/C

shelf. Vertical coverage at each site varies substantially
from acoustic Doppler current profilers (ADCPs) with
bins every meter or less over about 80% of the water
column to moorings with as few as two current meters
(Table 1). For sites with sufficient vertical coverage,
depth-averaged flows are estimated using a trapezoidal
integration and assuming the flow is vertically uniform
near the boundaries to extrapolate to the surface and
bottom. Results are similar if the velocity profile is ex-
trapolated linearly to the surface and bottom. For sites
where accurate estimates of the depth-average flow are
not possible, because of short records at some depths or
only a few instruments in the vertical, interior currents
are used to represent the along-shelf flow (Table 1).
To characterize the wind forcing, observations from
five National Data Buoy Center (NDBC) buoys, three
towers, and three coastal masts spanning the MAB and
southern flank of Georges Bank are analyzed (Fig. 1;
Table 2). The wind time series are 10–20 yr long, with
the exception of two coastal sites: Martha’s Vineyard
Coastal Observatory (MVCO) south of Cape Cod and
Tuckerton (TCK), New Jersey. The NDBC buoys are
located over the mid- to outer shelf. The Buzzards Bay
(BUZ) west of Cape Cod, Ambrose (ALS) in New
York Bight, and Chesapeake Bay (CHL) towers are
over the inner shelf. The MVCO and Tuckerton masts
are at the coast and the mast at the Army Corps Field
Research Facility (FRF) near Duck, North Carolina, is
at the end of a pier. Wind stresses are estimated from
wind velocities and sensor heights using a bulk formula
(Large and Pond 1981).
Historical hydrographic observations from the Na-
tional Oceanographic Data Center’s (NODC’s) World
Ocean Database 2001 archive of ship observations are
used to characterize the seasonal variation in the cross-
shelf structure of temperature, salinity, and density. The
observations were quality controlled and water depths
were determined using the National Geophysical Data
Center high-resolution bathymetry for the region
(Lentz et al. 2003). A total of 20 158 profiles over the
shelf (water depth ⱕ100 m) were extracted, excluding
profiles in Chesapeake Bay, Delaware Bay, Long Is-
land Sound, Buzzards Bay, and Nantucket Sound. Each
shelf profile was interpolated onto a 5-m vertical grid.
TABLE 1. Summary of annual harmonic analysis of 27 along-shelf current time series. Positions, water depth h, instrument depth z
when an interior velocity is used instead of the depth-averaged “da”flow, and length of time series are included. Orientation is relative
to true north and phase is zero on 1 Jan. Error bars are 95% confidence intervals. Major and minor axis error bars, and orientation and
phase error bars are the same.
Latitude
(°N)
Longitude
(°W) h(m) z(m)
Length
(days)
Major
(cm s
⫺1
)
Minor
(cm s
⫺1
)
Orientation
(°N) Phase (°)
40°51.8⬘67°33.5⬘76 da 204 3.0 ⫺1.1 ⫾4.0 55.3 353.3 ⫾95.3
40°58.1⬘67°19.2⬘75 da 827 2.1 0.7 ⫾1.8 119.0 217.9 ⫾56.5
40°52.1⬘67°24.3⬘81 45 1065 3.3 ⫺0.4 ⫾1.5 76.2 72.1 ⫾27.0
41°20.2⬘70°33.4⬘12 da 1577 3.0 0.1 ⫾1.2 96.4 180.0 ⫾22.8
41°19.1⬘70°34.2⬘18 da 276 2.6 0.1 ⫾3.2 87.7 24.0 ⫾70.2
41°15.2⬘70°35.5⬘28 da 372 4.7 0.2 ⫾1.5 100.9 200.1 ⫾17.8
40°41.6⬘70°8.6⬘46 32 348 3.8 0.6 ⫾2.7 144.1 201.1 ⫾41.7
40°30.0⬘70°12.5⬘66 32 424 0.9 ⫺0.1 ⫾2.7 164.1 256.9 ⫾174.0
40°20.6⬘70°16.1⬘88 32 393 1.7 1.0 ⫾2.7 48.6 194.8 ⫾158.9
40°12.9⬘70°18.2⬘105 59 393 3.5 ⫺0.7 ⫾2.7 65.0 199.1 ⫾47.6
40°28.0⬘70°54.7⬘80 da 379 3.8 ⫺0.3 ⫾2.8 76.9 49.0 ⫾43.2
40°35.0⬘70°27.5⬘64 da 302 2.5 0.8 ⫾3.2 120.8 244.7 ⫾84.3
40°29.5⬘70°30.3⬘70 da 310 3.9 0.6 ⫾3.2 109.8 258.4 ⫾49.7
40°23.0⬘70°32.6⬘86 da 310 3.7 0.7 ⫾3.2 98.1 284.8 ⫾52.3
40°28.5⬘70°20.1⬘70 50 310 3.8 ⫺0.3 ⫾3.2 94.3 246.5 ⫾48.4
40°34.2⬘72°18.5⬘49 da 473 2.6 ⫺0.8 ⫾1.3 52.5 145.2 ⫾32.4
40°25.3⬘72°8.2⬘59 da 213 1.7 ⫺0.0 ⫾1.3 129.9 5.9 ⫾42.7
40°11.1⬘72°0.2⬘65 da 335 2.9 ⫺0.3 ⫾1.3 68.0 1730.8 ⫾25.3
40°6.6⬘72°55.2⬘47 da 818 1.0 0.4 ⫾1.3 55.7 341.8 ⫾99.7
39°15.9⬘73°1.4⬘70 da 261 3.1 0.0 ⫾1.3 35.6 140.1 ⫾23.8
39°24.3⬘73°43.2⬘32 da 416 2.5 0.6 ⫾1.3 49.5 115.2 ⫾31.8
39°27.7⬘74°15.7⬘11 da 510 3.6 0.1 ⫾1.8 29.0 287.3 ⫾29.0
38°43.6⬘73°39.3⬘61 da 720 1.3 ⫺0.7 ⫾1.9 51.7 315.6 ⫾122.1
37°42.0⬘74°20.4⬘90 da 399 1.2 ⫺0.3 ⫾2.7 68.7 108.1 ⫾144.4
36°14.7⬘75°42.5⬘22 16 536 2.9 ⫺0.4 ⫾2.1 ⫺10.4 209.1 ⫾41.5
36°14.7⬘75°12.4⬘35 20 461 2.3 ⫺0.5 ⫾2.1 ⫺19.5 217.3 ⫾55.1
36°14.6⬘74°54.4⬘60 30 560 2.1 0.2 ⫾2.1 16.5 309.5 ⫾55.7
1488 JOURNAL OF PHYSICAL OCEANOGRAPHY VOLUME 38

Seasonal variations in near-bottom cross-isobath
density differences ⌬
␳
b
and temperature differences
⌬T
b
were estimated as follows using the historical hy-
drographic profiles from the NODC archive: First, the
average for each day of the year of all density (or tem-
perature) observations within 20 m of the bottom, in a
20-m water depth band centered on a given isobath,
were computed for 10-m isobath intervals from 10 to
120 m. Then a seasonal cycle was fit to the density
differences between adjacent isobath bins. Seasonal
variations in ⌬T
b
were also determined from a few con-
current mooring pairs over the New England and
southern MAB shelf.
b. Harmonic analysis and uncertainty estimates
Seasonal cycles of currents, wind stresses, and cross-
shelf temperature and density gradients are estimated
by fitting observations to a mean and annual harmonic:
y共t兲⫽y⫹asin
␻
t⫹bcos
␻
t,
where yis a time series, tis time in days, yis the time
average,
␻
⫽2
␲
/(365.25 day)
⫺1
is the annual frequency,
and aand bare coefficients of the harmonic fit. For
scalar time series results of the harmonic analysis can
be summarized as an amplitude and phase of the annual
cycle. For vector time series, the aand bcoefficients are
complex and the results are summarized in terms of an
ellipse characterized by the major and minor axes, the
orientation of the major axis, and the phase. The fol-
lowing analysis focuses on the depth-averaged along-
shelf flow. Phases are presented as the time of year
when the equatorward flow (direction of the mean
flow) is a maximum.
Uncertainties in aand bare determined using the
residual current variance in a low-frequency band
around the annual frequency
␻
and assuming Gaussian
statistics. Uncertainties in the corresponding amplitude
and phase or ellipse characteristics are then determined
through a linearized error analysis. This is the same
procedure used to estimate uncertainties in tidal con-
stituents (Pawlowicz et al. 2002).
The error analysis can be applied directly to the wind
stress time series because they are typically 10–20 yr
long. Individual current time series are too short to
accurately estimate the spectral variance of the residual
time series in the frequency band around the annual
frequency. Examination of the current spectra from
the few sites with longer time series indicates the spec-
tral density in the along-shelf component of the flow
is ⬃10
4
(cm s
⫺1
)
2
(cpd)
⫺1
. Therefore, uncertainties in
the current amplitudes and phases at all sites are esti-
mated assuming this value for the spectral density of
the residual flow in the frequency band around the an-
nual frequency. The current time series are also too
short (200 days to 3 yr) to determine whether there is a
consistent seasonal cycle over many years at individual
sites. Therefore, confidence in the generality of the re-
sults is based on the consistency of the patterns ob-
served for all the sites, and the results are only sugges-
tive with regard to a consistent seasonal cycle.
3. Seasonal variations
a. Wind stress
In winter, the MAB is in the westerly wind band
(Isemer and Hasse 1985) and the mean winds are rela-
tively strong toward the southeast. From spring to sum-
mer, the subtropical high over the North Atlantic
strengthens and moves northward, displacing the west-
erlies, resulting in weak northwestward mean winds in
summer. As a result, wind stresses in the MAB exhibit
a significant, spatially uniform annual cycle (Fig. 1 and
Table 2). The amplitude of the major axis seasonal
TABLE 2. Summary of annual harmonic analysis of wind stresses. Orientation is relative to true north and phase is zero on 1 Jan.
Error bars are 95% confidence intervals on estimates.
Station
Latitude
(°N)
Longitude
(°W)
Duration
(yr)
Major axis
(10
⫺2
Nm
⫺2
)
Minor axis
(10
⫺2
Nm
⫺2
)
Orientation
(from N)
Phase
(°)
44011 41°5.4⬘66°35.4⬘16.4 3.6 ⫾0.5 0.0 ⫾0.4 143.9 ⫾7.0 186.9 ⫾7.9
MVCO 41°20.2⬘70°33.4⬘5.1 1.1 ⫾0.4 ⫺0.2 ⫾0.4 141.0 ⫾20.9 191.0 ⫾20.2
BUZ 41°24.0⬘71°1.8⬘18.6 4.3 ⫾0.6 ⫺0.6 ⫾0.6 148.2 ⫾7.9 181.3 ⫾8.0
44008 40°30.0·⬘69°25.8⬘18.9 4.2 ⫾0.6 0.0 ⫾0.6 136.7 ⫾7.7 184.3 ⫾7.7
ALS 40°27.6⬘73°49.8⬘19.6 3.5 ⫾0.4 0.1 ⫾0.4 128.7 ⫾6.9 188.3 ⫾6.6
44025 40°15.0⬘73°10.2⬘17.6 3.9 ⫾0.4 0.0 ⫾0.4 141.6 ⫾5.6 182.5 ⫾6.1
TCK 39°27.7⬘74°15.7⬘2.1 1.3 ⫾0.5 ⫺0.4 ⫾0.4 115.2 ⫾21.0 196.2 ⫾23.9
44009 38°27.6⬘74°42.0⬘17.0 3.8 ⫾0.5 0.2 ⫾0.6 132.3 ⫾8.6 182.1 ⫾8.2
CHL 36°54.6⬘75°42.6⬘17.5 3.7 ⫾0.5 1.2 ⫾0.4 119.1 ⫾7.5 171.0 ⫾8.9
FRF 36°10.8⬘75°45.0⬘22.0 1.5 ⫾0.3 0.6 ⫾0.3 100.8 ⫾12.6 175.4 ⫾14.8
44014 36°34.8⬘74°50.4⬘11.8 3.2 ⫾0.5 0.6 ⫾0.5 135.3 ⫾9.5 182.8 ⫾9.5
JULY 2008 L ENTZ 1489

variation is 0.03–0.04 N m
⫺2
at all the sites except that
the three coastal masts have substantially smaller am-
plitudes, 0.01–0.015 N m
⫺2
(Fig. 2a). The amplitudes
are reduced at the coast because the mean and seasonal
winds are offshore so that there are topographic effects
such as flow separation and the surface roughness over
land is generally larger than over water (e.g., Vickers et
al. 2001). Thermodynamic adjustment in the marine
boundary layer may also be important (e.g., Austin and
Lentz 1999). The cross-shore distance over which the
amplitude of the seasonal cycle increases to the shelf
values is not known but is probably 10 km or less since
the Buzzards Bay, Ambrose, and Chesapeake Light
towers are all close to shore but have amplitudes similar
to the offshore buoys. Consequently, buoy winds are
used in place of the three coastal wind stations when
determining the wind-driven response of the circula-
tion.
Minor axis amplitudes are not significantly different
from zero except in the southern MAB (Fig. 1; Table
2). The seasonal variation in wind stress is polarized in
the southeast to northwest direction (orientation
⬃135°N; Fig. 2b), with the maximum southeastward
wind stress in early January (Fig. 2c) and maximum
northwestward wind stress in June. Both the phase and
orientation are essentially the same throughout the
MAB. The mean wind stress is toward the southeast
with a magnitude of 0.02–0.03 N m
⫺2
, so the seasonal
variation reinforces the mean wind stress in winter and
opposes the mean wind stress in summer. Since the
FIG. 2. Major axis (a) amplitude and (b) orientation and (c) phase of wind stress seasonal
ellipses as a function of along-shelf distance from northeastern Georges Bank (0 km) to Cape
Hatteras (⫺1400 km). Orientations are relative to true north and phases are date of maximum
southeastward wind stress. Error bars indicate 95% confidence interval for estimates.
1490 JOURNAL OF PHYSICAL OCEANOGRAPHY VOLUME 38

minor axis is approximately zero, there is not a signif-
icant seasonal variation in the northeast to southwest
wind stress component (orientation ⬃45°N). This has
important consequences for the seasonal variations in
the wind-driven circulation over the MAB shelf be-
cause the along-shelf flow is driven by the northeast–
southwest component of the wind stress over much of
the MAB shelf.
b. Depth-average (interior) currents
Seasonal variations in the depth-average (or interior)
flow are polarized at all sites (Figs. 1 and 3), with major
axis amplitudes of 1–5cms
⫺1
(Table 1). The ampli-
tudes are significantly different from zero at the 95%
confidence level for 18 of the 27 time series. The un-
certainties tend to be large relative to the amplitudes
because the time series are only about 1 yr long. The
orientation of the seasonal variation is generally
aligned with the principal axes of the subtidal flow,
which is along-isobath, though there are exceptions
(Figs. 3b and 1). Minor axis (cross-isobath) amplitudes
are all insignificant at the 95% confidence level (Table
1). The remaining analyses focus on the depth-average
along-shelf flow, where along-shelf is defined as the
orientation of the major axes of the subtidal flow, posi-
tive poleward. No consistent pattern was found for sea-
sonal variations in the cross-shelf component of flow at
any depth, probably because the uncertainties are large
relative to the amplitudes of the seasonal variations for
these short time series.
Amplitudes of the seasonal variation in the depth-
FIG. 3. (a) Ratio of minor to major axes of seasonal current ellipses and (b) seasonal ellipse
orientation (
␪
s
) relative to principal axes of subtidal flow (
␪
pa
) both as a function of signal-
to-noise (ratio of major axis to uncertainty in major axis).
JULY 2008 L ENTZ 1491

average along-shelf flow are 1–5cms
⫺1
with no clear
dependence on water depth or along-shelf position
(Fig. 4a). There are large variations in the time of maxi-
mum equatorward along-shelf flow (phase) with a ten-
dency for the phase to increase with increasing water
depth (Fig. 4b). The seasonal variations in the along-
shelf flow consist of at least two components: a wind-
driven component associated with seasonal variations
in the wind stress and a buoyancy-driven component
associated with seasonal variations in the cross-shelf
density gradient. Therefore, the wind-driven compo-
nent is examined first and then removed from the cur-
rent time series to isolate the buoyancy-driven compo-
nent.
c. Along-shelf current response to wind forcing
Along-shelf currents at all 27 sites are significantly
correlated with the wind stress (correlations 0.5–0.8).
However, there are substantial spatial variations in the
along-shelf current response to the wind stress forcing.
Both the amplitude of the response to the wind stress
(Noble et al. 1983) and the orientation of the wind
stress that is most correlated with the along-shelf cur-
rents (Beardsley et al. 1985; Shearman and Lentz 2003;
Kohut et al. 2004) vary within the MAB. Regression
coefficients between the wind stress and the subtidal
depth-averaged along-shelf flow increase by a factor of
2 or more from Georges Bank to the southern MAB
FIG. 4. (a) Major axis amplitude and (b) phase of along-shelf current from harmonic analysis
as a function of water depth. Phases are time (month) of maximum equatorward flow. Error
bars indicate 95% confidence interval for estimates. Minor axis amplitudes (not shown) are
not significantly different from zero for any of the time series (Table 1).
1492 JOURNAL OF PHYSICAL OCEANOGRAPHY VOLUME 38

(Fig. 5a). The wind stress orientation that yields maxi-
mum correlation with the along-shelf flow is approxi-
mately north–south (10°–20°N, solid square in Fig. 5b)
south of Chesapeake Bay, east–west (90°–110°N) near
the coast south of Martha’s Vineyard (solid circle: wa-
ter depths less than 60 m), and northeast–southwest
(30°–70°N) at all the other sites (Georges Bank, mid
and outer New England shelf, and the central MAB).
The wind-driven flow was estimated by rotating the
wind stress to the orientation of maximum correlation
(Fig. 5b) and multiplying by the regression coefficient
(Fig. 5a) for each site. The seasonal variation of the
estimated wind-driven flow was then calculated for
each site. The regression coefficient and orientation
based on subtidal values are used because regression
coefficients and orientations based on monthly values
are similar, but less certain, since there are fewer de-
grees of freedom. Monthly values of the estimated
wind-driven flow are generally significantly correlated
with monthly average along-shelf flows.
There is not a significant seasonal variation in the
wind-driven along-shelf flow at most sites (Fig. 6) be-
cause the component of the wind stress that drives the
depth-averaged along-shelf flow (40°–60°N in Fig. 5b)
is perpendicular to the orientation of the seasonal
variation in the wind stress (130°–150°N in Fig. 2b). In
FIG. 5. (a) Slope of linear regression as a function of along-shelf distance from northeastern
Georges Bank (0 km) to Cape Hatteras (⫺1400 km). Linear regression is
␷
da
⫽a
␶
s
⫹b, where
ais the slope and b(not shown) is the intercept, and
␶
s
is the component of the wind stress
that yields the maximum correlation with
␷
da
. (b) Orientation of the wind stress that yields the
maximum correlation with
␷
da
as a function of water depth. Dashed line indicates orientation
of wind stress component without a seasonal variation. Error bars in (a) indicate 95% confi-
dence interval.
JULY 2008 L ENTZ 1493

particular, in water depths greater than 60 m, the am-
plitudes of the seasonal variation in the estimated wind-
driven flow are 2 cm s
⫺1
or less (not shown) and are not
significantly different from zero at the 95% confidence
level.
In the southern MAB, where the regression slope is
large (solid square: Fig. 5a) and the orientation of the
wind stress that drives the depth-averaged along-shelf
flow is roughly north–south (solid square: Fig. 5b), the
wind drives a significant seasonal variation in the
depth-averaged along-shelf flow of about 3 cm s
⫺1
with
maximum southward along-shelf flow in winter (solid
square: Fig. 6).
Over the New England shelf, south of Cape Cod, the
orientation of the wind stress that drives the depth-
averaged along-shelf flow rotates from east–west over
the inner shelf to northeast–southwest over the middle
and outer shelf (solid circle: Fig. 5b). Consequently,
there is a significant seasonal variation of 2 cm s
⫺1
with
maximum westward along-shelf flow in summer over
the inner shelf that decreases offshore (Figs. 6 and 7).
Year-to-year variations in the monthly average currents
over the mid and outer shelf are large (Figs. 7c,d). This
is due in part to the impact of individual storm events
on the monthly values and emphasizes the need for
much longer time series to accurately determine sea-
sonal cycles in the circulation.
d. Current response related to cross-shelf buoyancy
gradients
Estimates of the subtidal wind-driven flow were sub-
tracted from the current time series at each site and the
seasonal harmonic analysis was applied to the residual
time series. Amplitudes of the seasonal harmonic of the
residual along-shelf flow range from 1 to 5 cm s
⫺1
and
only about half of the amplitudes are significantly dif-
ferent from zero (Fig. 8a). The amplitudes do not ex-
hibit any obvious systematic dependence on water
depth (offshore distance) or along-shelf position,
though any such pattern may be obscured by the large
uncertainties in the estimates.
The phases tend to fall into two groups (Fig. 8b).
Onshore of the 60-m isobath, maximum equatorward
flow tends to occur in the spring (February–June), par-
ticularly for sites west of 72°W (open symbols Fig. 8b).
This is consistent with the timing of maximum freshwa-
ter discharge from MAB estuaries (discussed below).
Offshore of the 60-m isobath maximum equatorward
flow tends to occur in the autumn (September–Decem-
ber). For sites south of Cape Cod (solid circle: Fig. 8b)
there is a steady phase increase with water depth from
May in shallow water to December–January near the
shelf break. The offshore fall maximum and the phase
increase with increasing water depth south of Cape Cod
FIG. 6. Amplitude of the seasonal cycle of the wind-driven depth-average along-shelf flow
as a function of along-shelf distance from northeastern Georges Bank (0 km) to Cape Hatteras
(⫺1400 km). Error bars indicate 95% confidence interval for estimates.
1494 JOURNAL OF PHYSICAL OCEANOGRAPHY VOLUME 38

are both consistent with the along-shelf flow being in
thermal wind balance with the cross-shelf density gra-
dient (Lentz et al. 1999; Shearman and Lentz 2003;
Garvine 2004; Codiga 2005) and the tendency for the
maximum near-bottom cross-isobath temperature dif-
ferences to occur later in the year in deeper water (dis-
cussed below).
There is a relatively strong cross-shelf salinity gradi-
ent over the MAB shelf with fresher water near the
coast (Fig. 9). However, the seasonal variations in both
salinity (Manning 1991; Mountain 2003) and the cross-
shelf salinity gradient (Shearman and Lentz 2003) tend
to be small, except over the inner shelf where there can
be an enhanced cross-shelf salinity gradient due to
spring runoff (March–April: Fig. 9) (Ullman and Co-
diga 2004; Codiga 2005). The maximum equatorward
flow in spring onshore of the 60-m isobath and west of
72°W (open symbols: Fig. 8b) may be associated with
spring runoff enhancing the cross-shelf density gradient
over the inner shelf. The Connecticut River is located at
about 72°W and has a well-documented impact on sea-
sonal variations in the flow in that area (Ullman and
FIG. 7. Monthly averages of the depth-average or interior along-shelf flow
␷
over the New
England shelf at (a) 12-m water depth, (b) 27-m (circle) and 46-m (triangle) water depth, (c)
⬃70-m water depth, and (d) ⬃85-m water depth. The estimated wind-driven along-shelf flow
for each site is also shown (dashed lines).
JULY 2008 L ENTZ 1495

Codiga 2004; Codiga 2005). The Hudson River, Dela-
ware Bay, and Chesapeake Bay all have peak freshwa-
ter discharges in spring. The associated buoyant coastal
currents reduce the inner-shelf salinity, increasing the
cross-shelf density gradient, and enhance the along-
shelf flow (e.g., Munchow and Garvine 1993; Rennie et
al. 1999; Yankovsky et al. 2000).
There is a large seasonal variation in the cross-shelf
temperature gradient (Fig. 10; Shearman and Lentz
2003). In winter (December–March), thermal stratifica-
tion is weak and the shelf water temperatures decrease
onshore because of surface cooling. In winter, the cross-
shelf temperature gradient (denser water near the
coast) opposes the cross-shelf salinity gradient (lighter
water near the coast), resulting in a reduced cross-shelf
density gradient. Surface heating in the spring leads to
the development of a shallow seasonal thermocline that
isolates a region of residual cold winter water near the
bottom over the mid and outer shelf, called the “cold
pool,”that persists from May through October (Fig. 10:
Bigelow 1933; Houghton et al. 1982). Note this is not
simply a local response. The cold-pool water is ad-
vected equatorward along-shelf by the mean flow and
hence comes from regions to the north (Houghton et al.
1982). Offshore of the center of the cold pool, the near-
bottom cross-shelf temperature gradient continues to
oppose the cross-shelf salinity gradient, as in winter.
However, onshore of the cold pool, the near-bottom
FIG. 8. (a) Amplitude and (b) phase of the seasonal ellipse of the residual depth-averaged
along-shelf flow when the wind-driven flow is removed as a function of water depth. The phase
of the near-bottom cross-shelf temperature or density difference is shown in (c). Error bars
indicate 95% confidence interval for estimates.
1496 JOURNAL OF PHYSICAL OCEANOGRAPHY VOLUME 38

cross-shelf temperature gradient enhances the cross-
shelf salinity gradient (warmer and fresher water near
the coast), resulting in an enhanced near-bottom cross-
shelf density gradient. The larger cross-shelf density
gradient implies an enhanced thermal wind shear and
therefore an enhanced along-shelf flow. The maximum
cross-shelf temperature gradient (temperature increas-
ing onshore) occurs in July over the inner shelf. As the
thermocline deepens, due to increasing wind events and
decreasing surface heat flux in the autumn (Beardsley
et al. 1985; Lentz et al. 2003), the maximum cross-shelf
temperature gradient moves offshore (cf. July and Sep-
tember: Fig. 10). The maximum cross-shelf density gra-
dient over the outer shelf, offshore of the center of the
cold pool, occurs in late autumn (November), the only
time of the year when the temperature does not in-
crease offshore over the outer shelf. As a consequence
of the seasonal evolution of the temperature field, the
maximum cross-shelf density gradient near the bottom
occurs in summer over the inner shelf, progressing to
the outer shelf in late autumn, consistent with the pro-
gression observed south of Cape Cod and the fall maxi-
mum offshore of the 60-m isobath.
4. Summary
Seasonal variations in the depth-average (or interior)
along-shelf flow are estimated for 27 sites with current
time series longer than 200 days over the continental
shelf of the Middle Atlantic Bight and southern flank of
FIG. 9. Cross-shelf sections of bimonthly mean salinities for the New England shelf
(69°–73°W).
JULY 2008 L ENTZ 1497

Georges Bank. Seasonal variations in the depth-
average along-shelf flow are polarized along-isobath
with major axis amplitudes of 1–5cms
⫺1
(Fig. 1). The
individual current time series are typically only about a
year in duration and, hence, are not long enough to
determine a reliable seasonal cycle. Nevertheless, the
seasonal variations from the 27 sites exhibit some con-
sistent patterns associated with seasonal variations in
the wind stress, surface heat flux, and spring river dis-
charge.
There is a significant, spatially uniform seasonal cycle
in the wind stress with an amplitude of about 0.03–0.04
Nm
⫺2
polarized in the southeast–northwest direction
(Fig. 2). However, the wind stress does not force a sig-
nificant seasonal variation in the along-shelf flow over
most of the MAB shelf and southern flank of Georges
Bank (Fig. 6) because the orientation of the wind stress
that forces the along-shelf flow is perpendicular to the
orientation of the seasonal variation in the wind stress
(Fig. 5b). There are significant seasonal variations in
the wind-driven flow in the southern MAB (amplitude
of3cms
⫺1
) and over the inner New England shelf
(amplitude of 2 cm s
⫺1
: see also Ullman and Codiga
2004; Codiga 2005) because there is a seasonal variation
in the component of the wind stress that forces the flow
in these regions.
When the wind-driven component of the flow is re-
moved, there is a seasonal variation (1–5cms
⫺1
)inthe
residual along-shelf flow with a phase such that the
maximum equatorward flow tends to occur in spring
onshore of the 60-m isobath and in autumn offshore of
the 60-m isobath (Fig. 8b). The phase variation is con-
FIG. 10. As in Fig. 9, but of monthly mean temperatures, showing the seasonal evolution.
1498 JOURNAL OF PHYSICAL OCEANOGRAPHY VOLUME 38

sistent with seasonal variations in the cross-shelf den-
sity gradient due to both salinity and temperature gra-
dients. The maximum equatorward flow in spring on-
shore of the 60-m isobath is consistent with the
maximum freshwater discharge from rivers and estuar-
ies occurring in spring. The maximum equatorward
flow in autumn offshore of the 60-m isobath and the
steady phase increase with water depth offshore of
Cape Cod are consistent with the seasonal variation in
the near-bottom cross-shelf temperature (density) gra-
dient. The seasonal variation in the cross-shelf tem-
perature gradient is, in turn, associated with the devel-
opment and destruction of a near-bottom pool of cold
water over the mid and outer shelf (“cold pool”) due to
surface heating in spring–summer and surface cooling
and wind-driven mixing in autumn–winter.
The seasonal variations in the depth-averaged along-
shelf flow over the MAB shelf exhibit reasonable pat-
terns that are sensible in the context of the seasonal
variations in wind stress, surface heat flux, and river
runoff. The sparse coverage and the large uncertainties
in the seasonal variations emphasize the need for much
longer time series with better spatial coverage to obtain
a more complete picture of the seasonal variations in
the circulation.
Acknowledgments. The author is grateful to the re-
searchers (too numerous to list) who collected the his-
torical data used here. The consistent pattern of weak
seasonal variations in the circulation that emerged from
this study is a testament to the care and effort that went
into collecting these observations. The author also ap-
preciates comments and suggestions on an early draft of
this manuscript by Bob Beardsley, Ken Brink, and
Melanie Fewings. The field programs were funded by
Department of Energy, Minerals Management Service,
National Science Foundation, National Oceanic and
Atmospheric Administration, and the Office of Naval
Research. This research was funded by the Ocean Sci-
ences Division of the National Science Foundation un-
der Grants OCE-820773, OCE-841292, and OCE-
848961.
REFERENCES
Aikman, F., III, H. W. Ou, and R. W. Houghton, 1988: Current
variability across the New England continental shelf-break
and slope. Cont. Shelf Res., 8, 625–651.
Austin, J. A., and S. J. Lentz, 1999: The relationship between syn-
optic weather systems and meteorological forcing on the
North Carolina inner shelf. J. Geophys. Res., 104, 18 159–
18 186.
Beardsley, R. C., and W. C. Boicourt, 1981: On estuarine and con-
tinental-shelf circulation in the Middle Atlantic Bight. Evo-
lution of Physical Oceanography: Scientific Surveys in Honor
of Henry Stommel, B. A. Warren and C. Wunsch, Eds., The
MIT Press, 198–233.
——, D. C. Chapman, K. H. Brink, S. R. Ramp, and R. Schlitz,
1985: The Nantucket Shoals Flux Experiment (NSFE79).
Part I: A basic description of the current and temperature
variability. J. Phys. Oceanogr., 15, 713–748.
Bigelow, H. B., 1933: Studies of the waters on the continental
shelf, Cape Cod to Chesapeake Bay, I, The cycle of tempera-
ture. Pap. Phys. Oceanogr. Meteor., 2, 1–135.
——, and M. Sears, 1935: Studies of the waters on the continental
shelf, Cape Cod to Chesapeake Bay, II, Salinity. Pap. Phys.
Oceanogr. Meteor., 4, 1–94.
Brink, K. H., R. Limeburner, and R. C. Beardsley, 2003: Proper-
ties of flow and pressure over Georges Bank as observed with
near-surface drifters. J. Geophys. Res., 108, 8001, doi:10.1029/
2001JC001019.
Bumpus, D. F., 1973: A description of the circulation on the con-
tinental shelf of the east coast of the United States. Prog.
Oceanogr., 6, 111–157.
Butman, B., 1987: Processes causing surficial sediment movement.
Georges Bank, R. H. Backus, Ed., The MIT Press, 147–162.
——, and R. C. Beardsley, 1987: Long-term observations on the
southern flank of Georges Bank. Part I: A description of the
seasonal cycle of currents, temperature, stratification, and
wind stress. J. Phys. Oceanogr., 17, 367–384.
Codiga, D. L., 2005: Interplay of wind forcing and buoyant dis-
charge off Montauk Point: Seasonal changes to velocity struc-
ture and a coastal front. J. Phys. Oceanogr., 35, 1068–1085.
Flagg, C. N., 1987: Hydrographic structure and variability.
Georges Bank, R. H. Backus, Ed., The MIT Press, 108–124.
——, and M. Dunn, 2003: Characterization of the mean and sea-
sonal flow regime on Georges Bank from shipboard acoustic
Doppler current profiler data. J. Geophys. Res., 108, 8002,
doi:10.1029/2001JC001257.
——,——, D.-P. Wang, H. T. Rossby, and R. L. Benway, 2006: A
study of the currents of the outer shelf and upper slope from
a decade of shipboard ADCP observations in the Middle
Atlantic Bight. J. Geophys. Res., 111, C06003, doi:10.1029/
2005JC003116.
Garvine, R. W., 2004: The vertical structure and subtidal dynam-
ics of the inner shelf off New Jersey. J. Mar. Res., 62, 337–371.
Houghton, R. W., R. Schlitz, R. C. Beardsley, B. Butman, and
J. L. Chamberlin, 1982: The Middle Atlantic Bight cold pool:
Evolution of the temperature structure during summer, 1979.
J. Phys. Oceanogr., 12, 1019–1029.
Isemer, H.-J., and L. Hasse, 1985: Observations. Vol. 1, The Bun-
ker Climate Atlas of the North Atlantic Ocean, Springer-
Verlag, 218 pp.
Kohut, J. T., S. M. Glenn, and R. J. Chant, 2004: Seasonal current
variability on the New Jersey inner shelf. J. Geophys. Res.,
109, C07S07, doi:10.1029/2003JC001963.
Large, W. G., and S. Pond, 1981: Open ocean momentum flux
measurements in moderate to strong winds. J. Phys. Ocean-
ogr., 11, 324–336.
Lentz, S. J., 2008: Observations and a model of the mean circula-
tion over the Middle Atlantic Bight continental shelf. J. Phys.
Oceanogr., 38, 1203–1221.
——, R. T. Guza, S. Elgar, F. Feddersen, and T. H. C. Herbers,
1999: Momentum balances on the North Carolina inner shelf.
J. Geophys. Res., 104, 18 205–18 226.
——, K. Shearman, S. Anderson, A. Plueddemann, and J. Edson,
2003: The evolution of stratification over the New England
shelf during the Coastal Mixing and Optics study, August
JULY 2008 L ENTZ 1499

1996–June 1997. J. Geophys. Res., 108, 3008, doi:10.1029/
2001JC001121.
Linder, C. A., and G. Gawarkiewicz, 1998: A climatology of the
shelfbreak front in the Middle Atlantic Bight. J. Geophys.
Res., 103, 18 405–18 423.
Manning, J., 1991: Middle Atlantic Bight salinity: Interannual
variability. Cont. Shelf Res., 11, 123–137.
Mayer, D. A., D. V. Hansen, and D. A. Ortman, 1979: Long-term
current and temperature observation on the Middle Atlantic
shelf. J. Geophys. Res., 84, 1776–1792.
Mountain, D. G., 2003: Variability in the properties of shelf water
in the Middle Atlantic Bight, 1977–1999. J. Geophys. Res.,
108, 3014, doi:10.1029//2001JC001044.
Munchow, A., and R. W. Garvine, 1993: Dynamical properties of
a buoyancy-driven coastal current. J. Geophys. Res., 98,
20 063–20 078.
Noble, M., B. Butman, and E. Williams, 1983: On the longshore
structure and dynamics of subtidal currents on the eastern
United States continental shelf. J. Phys. Oceanogr., 13, 2125–
2147.
Pawlowicz, R., R. Beardsley, and S. J. Lentz, 2002: Classical tidal
harmonic analysis with errors in MATLAB using T_TIDE.
Comput. Geosci., 28, 929–937.
Rennie, S., J. L. Largier, and S. J. Lentz, 1999: Observations of
low-salinity coastal current pulses downstream of Chesa-
peake Bay. J. Geophys. Res., 104, 18 227–18 240.
Saunders, P. M., 1977: Wind stress on the ocean over the eastern
continental shelf of North America. J. Phys. Oceanogr., 7,
555–566.
Shearman, R. K., and S. J. Lentz, 2003: Dynamics of mean and
subtidal flow on the New England shelf. J. Geophys. Res.,
108, 3281, doi:10.1029/2002JC001417.
Ullman, D. S., and D. L. Codiga, 2004: Seasonal variation of a
coastal jet in the Long Island Sound outflow region based on
HF radar and Doppler current observations. J. Geophys.
Res., 109, C07S06, doi:10.1029/2002JC001660.
Vickers, D., L. Mahrt, J. Sun, and T. Crawford, 2001: Structure of
offshore flow. Mon. Wea. Rev., 129, 1251–1258.
Yankovsky, A. E., R. W. Garvine, and A. Munchow, 2000: Meso-
scale currents on the inner New Jersey shelf driven by the
interaction of buoyancy and wind forcing. J. Phys. Oceanogr.,
30, 2214–2230.
1500 JOURNAL OF PHYSICAL OCEANOGRAPHY VOLUME 38