Diagnosing the warming of the Northeastern U.S. Coastal Ocean in 2012: A linkage between the atmospheric jet stream variability and ocean response

Abstract

The temperature in the coastal ocean off the northeastern U.S. during the first half of 2012 was anomalously warm, and this resulted in major impacts on the marine ecosystem and commercial fisheries. Understanding the spatio-temporal characteristics of the warming and its underlying dynamical processes is important for improving ecosystem management. Here we show that the warming in the first half of 2012 was systematic from the Gulf of Maine to Cape Hatteras. Moreover, the warm anomalies extended through the water column, and the local temperature change of shelf water in the Middle Atlantic Bight was largely balanced by the atmospheric heat flux. The anomalous atmospheric jet stream position induced smaller heat loss from the ocean and caused a much slower cooling rate in late autumn and early winter of 2011-2012. Strong jet stream intraseasonal oscillations in the first half of 2012 systematically increased the warm anomalies over the continental shelf. Despite the importance of advection in prior Northeast U.S. continental shelf inter-annual temperature anomalies, our analyses show that much of the 2012 warming event was attributed to local warming from the atmosphere.

Full text

Diagnosing the warming of the Northeastern U.S. Coastal Ocean in
2012: A linkage between the atmospheric jet stream variability and
ocean response
Ke Chen,
1
Glen G. Gawarkiewicz,
1
Steven J. Lentz,
1
and John M. Bane
2
Received 30 August 2013; revised 3 December 2013; accepted 4 December 2013.
[1] The temperature in the coastal ocean off the northeastern U.S. during the first half of
2012 was anomalously warm, and this resulted in major impacts on the marine ecosystem
and commercial fisheries. Understanding the spatiotemporal characteristics of the warming
and its underlying dynamical processes is important for improving ecosystem management.
Here, we show that the warming in the first half of 2012 was systematic from the Gulf of
Maine to Cape Hatteras. Moreover, the warm anomalies extended through the water
column, and the local temperature change of shelf water in the Middle Atlantic Bight was
largely balanced by the atmospheric heat flux. The anomalous atmospheric jet stream
position induced smaller heat loss from the ocean and caused a much slower cooling rate in
late autumn and early winter of 2011–2012. Strong jet stream intraseasonal oscillations in
the first half of 2012 systematically increased the warm anomalies over the continental
shelf. Despite the importance of advection in prior northeastern U.S. continental shelf
interannual temperature anomalies, our analyses show that much of the 2012 warming event
was attributed to local warming from the atmosphere.
Citation: Chen, K., G. G. Gawarkiewicz, S. J. Lentz, and J. M. Bane (2014), Diagnosing the warming of the Northeastern U.S. Coastal
Ocean in 2012: A linkage between the atmospheric jet stream variability and ocean response, J. Geophys. Res. Oceans,119,
doi:10.1002/2013JC009393.
1. Introduction
[2] The coastal ocean off the northeastern U.S. (Figure 1),
which encompasses the Gulf of Maine (GoM) and the Mid-
dle Atlantic Bight (MAB), lies in the confluent zone of the
North Atlantic’s subtropical and subpolar gyres and thus is
subject to influences from both the Gulf Stream and the
Labrador Current. The general circulation over the conti-
nental shelf is equatorward flow with complex shelf-deep
ocean exchanges across the shelf break [Loder et al.,
1998]. The MAB shelf break and Georges Bank are also
distinguished by high biological productivity with econom-
ically important commercial fisheries [e.g., Marra et al.,
1990; Ryan et al., 1999; O’Reilly et al., 1987]. Therefore,
understanding the ecosystem dynamics and its response to
the physical environment in this region have long been rec-
ognized as important [e.g., Fogarty and Murawski, 1998;
He et al., 2011; Ji et al., 2008].
[3] The coastal ocean in the MAB and GoM was anoma-
lously warm in the first half of 2012, which led to major
impacts on the northeastern U.S. coastal ecosystem and
commercial fisheries. According to a National Oceanic and
Atmospheric Administration (NOAA) Northeast Fishery
Science Center ecosystem advisory (issued in August
2012) [Friedland, 2012], the surface temperature of the
northeastern continental shelf during the first half of 2012
was the highest on record based on both contemporary sat-
ellite remote sensing data and long-term ship-board meas-
urements from the past 150 years ; The spring bloom in
2012 started as early as February, and phytoplankton bio-
mass was higher than average ; The distribution of Atlantic
Cod had a northward shift consistent with the response to
the warming over the continental shelf. From the CINAR
(Cooperative Institute for the North Atlantic Region) Shelf
break Ecosystem Workshop in January 2013 [Gawarkie-
wicz et al., 2013], commercial fishermen reported an abun-
dance of squid in the summer of 2012 as well as the
appearance of warm water species not previously seen off
southern New England. Implications of this dramatic
warming for management of living marine resources are
discussed in Mills et al. [2013]. Thus, it is of great impor-
tance to understand the spatial and temporal characteristics
of the warming and its underlying dynamical processes.
[4] Long-term sea surface temperature (SST) changes in
the coastal ocean off the northeastern U.S. are mainly con-
trolled by along-shelf transport originating in the Labrador
Sea [Shearman and Lentz, 2010]. On the U.S. west coast, the
atmospheric jet stream has also been reported to have major
impacts on the temperature as well as ecosystem changes
in the coastal ocean on an intraseasonal scale (20 day)
1
Department of Physical Oceanography, Woods Hole Oceanographic
Institution, Woods Hole, Massachusetts, USA.
2
Department of Marine Sciences, University of North Carolina, Chapel
Hill, North Carolina, USA.
Corresponding author: K. Chen, Woods Hole Oceanographic Institu-
tion, Physical Oceanography, 266 Woods Hole Rd., Woods Hole, MA
02543, USA. (kchen@whoi.edu)
©2013. American Geophysical Union. All Rights Reserved.
2169-9275/14/10.1002/2013JC009393
1
JOURNAL OF GEOPHYSICAL RESEARCH: OCEANS, VOL. 119, 1–10, doi:10.1002/2013JC009393, 2014

[Bane et al., 2007]. On a seasonal time scale, the shift of
the jet stream’s latitude is also associated with the timing
of initiation of upwelling favorable wind patterns on the
west coast of the U.S. [Barth et al., 2007].The processes
which determine the temperature anomalies in the coastal
ocean off the northeastern U.S. on a seasonal scale are not
well known at present. Here, we aim to understand the spa-
tiotemporal characteristics of the warm anomalies and
characterize the relative importance of ocean advection and
atmospheric forcing to the temperature budget in the
coastal ocean in the MAB during the 2012 warming event.
However, the lack of depth-dependent measurements dur-
ing this event inhibits our effort to fully close the tempera-
ture budget. Therefore, we examine the contribution of
atmospheric heat flux and attribute the residual to ocean
advective flux. Improved understanding of the dynamical
processes controlling the continental shelf warming will
lead to better ecosystem management and forecasting.
2. Method and Data
[5] We examined surface air and water temperature data
from National Data Buoy Center (NDBC) buoys within the
MAB and GoM. We chose four buoys (Figure 1) that are
located in the GoM (44005), and on the continental shelf
near Nantucket (44008), Long Island (44025), and Virginia
Beach (44099). For comparison with recent seasonal aver-
ages, water temperature records from 2000 to 2010 serve as
temporal means to compare against temperature in 2011
and 2012 at each buoy. At buoy 44099, water temperature
measurements are only available from 2008 to present. To
construct a continuous time series, water temperature data
from nearby buoy cbbv2 (near the Chesapeake Bay Bridge
Tunnel) and chlv2 (Chesapeake Light, the same location as
44099) are used. The temperature record during 2000–2004
at buoy chlv2 was used in place of 44099. Water tempera-
ture data during 2005–2007 was projected for buoy 44099
based on linear regression between temperature records
during 2008–2011 at buoy 44099 and cbbv2. All time
series in this study were 8 day low-pass filtered to remove
weather-band signals.
[6] To extend the analysis from the SST to the water col-
umn over the continental shelf, expendable bathythermo-
graph (XBT) data collected off New Jersey along the
Oleander line (from New York Harbor to Bermuda) were
examined. The XBT transects are available on a monthly
basis since 1977 (po.msrc.sunysb.edu/Oleander/XBT/
NOAA_XBT.html). We used the vertical profile closest to
Figure 1. Map of the Middle Atlantic Bight (MAB) and Gulf of Maine (GoM) showing locations of
NDBC buoys (diamonds) and the Oleander line (black dashed line). Shown in the background is the
mean SST anomaly in March 2012 referenced to the March average SST during 2000–2010. White con-
tours represent 1 and 2 standard deviations. Square boxes (0.25) demonstrate the location of WOA
profiles which are used to estimate the relationship between SST and depth-averaged temperature for
nearby buoys. The 50, 100, 200, and 1000 m isobaths (ETOPO-1) are also shown. The thin black line is
the smoothed 200 m isobath that is used to define along-shelf and cross-shelf directions.
CHEN ET AL.: DIAGNOSING THE COASTAL WARMING IN 2012
2

the 100 m isobath for each month, which was representa-
tive of the thermal structure over the middle and outer con-
tinental shelf.
[7] Atmospheric data are available from the National
Centers for Environmental Prediction (NCEP) North Amer-
ican Regional Reanalysis (NARR). The NARR data set has
temporal and spatial resolution of 3 h and 30 km, respec-
tively, and it spans from 1979 to the present. Shortwave
radiation, longwave radiation, latent heat flux, and sensible
heat flux components are used to calculate net downward
heat flux into the ocean. To determine the latitude of the
atmospheric jet stream, the maximum meridional gradient
of geopotential height at 200 hPa along 70W was used
[Bane et al., 2007]. Wind speed and specific humidity at 10
m above the sea surface are also available in the NARR
data set.
[8] At the three buoy locations in the MAB, the depth-
averaged temperature was estimated based on the closest
grid point (square boxes, Figure 1) of the World Ocean
Atlas (WOA) 2009 climatology (www.nodc.noaa.gov/
OC5/WOA09/pr_woa09.html). The climatological data
were used to develop a linear relationship between the cli-
matological surface temperature and climatological depth-
averaged temperature. The linear relationship was then
used to estimate depth-averaged temperature at buoy loca-
tions from the surface temperatures during 2011–2012. At
each location, we further produced two more profiles, i.e.,
61 standard deviation (from WOA09) of the original pro-
file, to account for the interannual variability of the thermal
structures. Thus two more linear coefficients were calcu-
lated at each location and were used to estimate depth-
averaged temperature in order to provide the range of
uncertainties.
[9] Daily Optimal Interpolated (OI) Advanced Very
High Resolution Radiometer (AVHRR) SST fields from
the National Climate Data Center (NCDC) were used to
compute the SST anomaly for March 2012 (Figure 1). The
data are available from September 1981 to the present with
a spatial resolution of 0.25. The SST anomaly in March
2012 was computed relative to the mean SST in March dur-
ing 2000–2010.
3. Results
3.1. Buoy Temperatures
[10] SST records at four NDBC buoys during 2011–2012
are compared to the mean values from 2000 to 2010 in Fig-
ure 2. The shelf-wide warm anomalies observed in the first
half of 2012 began in fall 2011. Systematic anomalous SST
was observed from November 2011 to at least June 2012.
The warm anomalies in the western GoM (44005) and on
the continental shelf near Nantucket (44008), Long Island
(44025), and Virginia Beach (44099) varied from 0 to 6C,
with the maximum anomaly of 6C near Virginia Beach in
March 2012. The average SST anomalies during the first
half of 2012 (January to June) are 1.7C, 1.9C, 2.4C, and
2.2C, respectively, from the Gulf of Maine to Virginia
Beach. While the average magnitudes of the warm anoma-
lies are not extremely large, the much larger short-term
fluctuations (up to 6C), the large area of the anomalies
(Figure 1), and the long duration of the shelf-wide warm
anomalies (at least 8 months) are significant and affected a
large number of marine organisms.
[11] The timing of the seasonal-scale warm anomalies
(November 2011 to June 2012) is largely consistent from
the GoM to the MAB. The short-term anomalies at 44008
(Nantucket) during early October and late November of
2011 are the results of a northward diversion of the Gulf
Stream [Gawarkiewicz et al., 2012]. On an intraseasonal
time scale, warm anomalies are in-phase from north to
south, particularly during March to May 2012. Calculation
of cross correlation between buoys (Table 1) reveals that
SST anomalies are highly correlated with maximum corre-
lation at zero lag. Further comparison also shows that cor-
relations in 2012 are higher than other years during 2000–
2010. The consistent timing of warm anomalies suggests
that the warming is related to large-scale atmospheric forc-
ing as opposed to along-shelf advection.
3.2. Vertical Structure of the Thermal Anomalies
[12] To determine the vertical extent of the warm
anomalies, we examine Oleander XBT temperature profiles
near the shelf break off New Jersey (Figure 3). The mean
profiles from 2000 to 2010 for each month are representa-
tive of the mean thermal structure on the continental shelf.
In comparison, monthly profiles in the first half of 2012
show that the warm anomalies extended from the surface
through the water column to at least 50 m depth near the
shelf break. During January to April when the upper water
column was well mixed, warm anomalies ranged from 1 to
3C from the surface to 50 m. In May 2012, the tempera-
ture in the upper 15 m was 5C warmer than the mean,
and the warm anomaly decreased with depth and reached
zero below 30 m. In June 2012, although the thermal strati-
fication remained similar to the mean condition, the tem-
perature was about 2–3C warmer over the upper 50 m of
the water column. The magnitudes of the XBT temperature
anomalies at the surface are comparable to those of the
Long Island buoy measurements, despite the spatial
separation.
3.3. Heat Budget for 2011–2012
[13] The depth-averaged heat budget can be approxi-
mated as:
@T
@t
5Q
qocpH
2ð
0
2H
urT0dz (1)
where T is the depth-averaged temperature ; q
0
is the aver-
age seawater density (1024 Kg m
23
); c
p
is the specific heat of
seawater (4190 J Kg
21
C
21
); H is the local water depth; Q is
the net atmospheric heat flux; and uand T0are the
depth-dependent velocity vector and temperature,
respectively.
[14] Due to the temporal similarities of the temperature
anomalies throughout the MAB, it is reasonable to assume
the temperature anomalies in the first half of 2012 are
largely due to the atmospheric forcing. Thus, we examine
the balance of the local rate of change of depth-averaged
temperature and the net atmospheric heat flux, and attribute
the residual to oceanic advective flux. To convert surface
temperature T
s
to depth-averaged temperature T at the
CHEN ET AL.: DIAGNOSING THE COASTAL WARMING IN 2012
3

three buoy locations in the MAB, we utilize the nearby
monthly climatology profiles from WOA 2009 (Figure 1)
to compute the linear relationship between surface temper-
ature and depth-averaged temperature. Monthly linear coef-
ficients between T
s
and T are then interpolated to each day
and are used to generate daily time series of T.
[15] Following Lentz et al. [2003], integration of (1) in
time gives:
T2T05Qcum 2Qadv (2)
where T
0
is the initial condition of T,
Qcum 5ðt
0
Q
qocpHdt0(3)
is the cumulative atmospheric net heat flux, and
Qadv 5ðt
0ð
0
-H
urT0dz
0
@1
Adt0
is the advective heat flux.
[16]T2T
0
and Q
cum
are then examined to characterize
the relative importance of atmospheric forcing and ocean
advection to the depth-averaged temperature budget at the
three buoys in the MAB. During November 2011 to June
Figure 2. (a–d) SST during late 2011 and 2012 (red) compared to the 2000–2010 mean (blue) and
standard deviation (shaded). The four buoys (from top to bottom) are located in the GoM (44005), and
on the continental shelf near Nantucket (44008), Long Island (44025), and Virginia Beach (44099). (e–h)
SST anomalies with respect to 2000–2010 mean at the same four buoys. The shelf-wide systematic
warming period is demonstrated in the shaded area.
Table 1. Maximum Cross-Correlation Coefficients Among Tem-
perature Anomalies (Daily, 8 Day Low-Pass Filtered) in the First
Half of 2012
a
Gulf of
Maine
(44005)
Nantucket
Shoals
(44008)
Long
Island
(44025)
Virginia
Beach
(44099)
Gulf of Maine (44005) 1 (0) 0.94 (0) 0.93 (0) 0.85 (0)
Nantucket Shoals
(44008)
0.94 (0) 1 (0) 0.97 (0) 0.93 (0)
Long Island (44025) 0.93 (0) 0.97 (0) 1 (0) 0.91 (0)
Virginia Beach (44099) 0.83 (0) 0.93 (0) 0.91 (0) 1 (0)
a
Lags (in days) corresponding to the maximum coefficients are shown
in parentheses.
CHEN ET AL.: DIAGNOSING THE COASTAL WARMING IN 2012
4

2012, the temperature budget of shelf water near either
Nantucket (buoy 44008) or Long Island (buoy 44025) is
largely balanced between the atmospheric heat flux and the
local rate of change (Figure 4), consistent with previous
studies at the southern flank of Georges Bank [Brink et al.,
2009]. However, at Virginia Beach (buoy 44099), the rela-
tionship between T2T
0
and Q
cum
is relatively weak. Pre-
sumably, the proximity of the buoy location to the Gulf
Stream and major estuaries leads to a significant horizontal
heat flux due to both onshore intrusions of Gulf Stream and
large temperature gradients, and these thus counterbalance
the Q
cum
. Considering the local water depth is about 200 m,
the heat budget at the Gulf of Maine buoy (44005) is
unlikely to have temperature changes balanced by the sur-
face flux, and thus is excluded from analysis.
3.4. The Jet Stream Position and Variability During
the First Half of 2012
[17] Buoy temperature records during March 2012 to May
2012 show in-phase oscillations of surface temperature from
the GoM to Virginia Beach with a period of 20 days (Fig-
ure 2), which is consistent with the typical periodicity of
intraseasonal oscillations of the atmospheric jet stream
[Bane et al., 2007; Barth et al., 2007]. The cause of the
intraseasonal oscillations of the jet stream is normally attrib-
uted to the mountain torques [e.g., Lott et al., 2004a, 2004b],
but detailed discussion of this process is out of the scope of
current study. Comparison between the SST anomalies and
the jet stream latitude anomaly confirms that the surface
temperature oscillation is related to north–south shifts of the
jet stream’s position (Figure 5). Correlation coefficients
between the latitudinal anomaly of the jet stream and the
SST anomalies during the first half of 2012 vary between
0.45 and 0.64 with lags of 3–6 days at different buoy loca-
tions. In particular, a higher correlation is found at all four
buoys during March–May with increased coefficients vary-
ing from 0.55 to 0.84. Although the degrees of freedom
[e.g., Emery and Thomson, 2004] are small during this 3
month period (8–10), the correlation is significant (95%)
at Nantucket and Virginia Beach. Around early March 2012,
the net atmospheric heat flux turns positive (into the ocean)
and increases water column stratification. Compared to the
weakly stratified winter season (i.e., January–February),
atmospheric forcing would have a more significant impact
on the surface temperature as most of the atmospheric heat
flux is limited to the surface mixed layer. That presumably
explains the higher correlation between the jet stream posi-
tion and the SST anomalies in the spring.
3.5. Atmospheric Forcing and the Jet Stream Position
[18] The intraseasonal oscillations of the jet stream have
a clear impact on the SST oscillations in the GoM and the
5 10 15
−50
−40
−30
−20
−10
0a)Jan
depth [m]
5 10 15
−50
−40
−30
−20
−10
0b)Feb
5 10 15
−50
−40
−30
−20
−10
0c)Mar
5 10 15
−50
−40
−30
−20
−10
0d)Apr
depth [m]
T [de
g
.C]
5 10 15
−50
−40
−30
−20
−10
0e)May
T [de
g
.C]
5 10 15
−50
−40
−30
−20
−10
0f) Jun
T [de
g
.C]
2000−2010 mean
2000−2010 SE
2012
Figure 3. Comparison of Oleander XBT temperature profiles between 2012 (red) and 2000–2010
mean (blue) with standard errors shown by dashed lines. Profiles that are located closest to the 100 m iso-
bath in each month are selected. To avoid near-bottom variability associated with shelf break frontal
motions, only the upper 50 m are shown.
CHEN ET AL.: DIAGNOSING THE COASTAL WARMING IN 2012
5

MAB. In general, the atmospheric forcing of the upper
ocean includes both the net heat flux, net fresh water flux,
and the wind stress. All three are potentially affected by
meridional shifts in the position of the jet stream. During
the late fall and winter of 2011–2012 (November 2011 to
February 2012), the net atmospheric heat flux (downward
into the ocean) over the GoM and the MAB was anomalously
positive with a peak value up to 221 W m
22
(Figure 6),
compared to the mean heat fluxes (about 2200 to 2100 W
m
22
during November to February) during 2000–2010. The
positive heat flux anomaly was closely connected to the
anomalous northward shift of the jet stream, oscillating at a
period of about 20–30 days. During this time, the shortwave
and longwave radiation values were similar to the 2000–
2010 means. However, the latent and sensible heat fluxes
(downward into the ocean) were approximately 100 W m
22
larger, indicating the ocean was gaining more heat than usual.
According to bulk flux formulas [e.g., Fairalletal., 1996],
the upward latent heat flux, Q
L
, and sensible heat flux, Q
S
,
are defined as :
QL5qaLECLU10 qs2qa
ðÞ (4)
QS5qaCa
pCSU10 Ts2Ta
ðÞ (5)
where q
a
is the air density (1.3 Kg m
23
); L
E
is the latent
heat of evaporation (2.5 310
6
JKg
21
); C
L
is the latent
heat transfer coefficient (1.2 310
23
); U
10
is the wind
speed at 10 m above the sea surface ; q
s
and q
a
are specific
humidity of air at the sea surface and 10 m above the surface;
Ca
pis the specific heat capacity of air (1030 J Kg
21
C
21
);
C
S
is the sensible heat transfer coefficient (1.0 310
23
); and
T
s
and T
a
are temperature at the sea surface and air tem-
perature at 10 m above the sea surface. Analysis of U
10
,
T
a
and q
a
reveals that the wind speed was relatively weak
during this cooling period (November 2011 to February
2012) while the air humidity and temperature were rela-
tively higher than the mean condition (not shown). These
anomalies decrease the latent and sensible heat fluxes
from the ocean to the atmosphere and thus inhibit the total
heat loss of the ocean. Indeed, at all four buoys from the
GoM to Virginia Beach (44005, 44008, 44025, and
44099), the cooling rates during November 2011 to Febru-
ary 2012 were, respectively, 0.06Cd
21
,0.08
Cd
21
,
0.07Cd
21
, and 0.07Cd
21
, systematically smaller than
2000–2010 mean values of 0.07Cd
21
,0.09
Cd
21
,
0.10Cd
21
, and 0.10Cd
21
. Therefore, the anomalous
atmospheric heat flux increased the heat content in the
coastal ocean during the preceding winter of 2012. During
March 2012 to May 2012, strong jet stream intraseasonal
oscillations also induced a positive anomalous atmos-
pheric heat flux, though not as large as that during the
cooling period. With the development of stratification, the
warming was not distributed evenly through the water col-
umn, which occurred during the winter, but was concen-
trated above the seasonal thermocline. Thus, the smaller
heat flux anomalies could still drive large SST anomalies.
N D J F M A M J J A S O N D J
−20
−10
0
10
20
2011 2012 2013
deg.C
44099
T−T0
Qcum
−20
−10
0
10
20
deg.C
44025
−20
−10
0
10
20
deg.C
44008
Figure 4. Depth-averaged temperature (solid) and cumulative heat flux (dashed) during late 2011 and
2012 at three buoys in the MAB. Gray curves represent the error range associated with the estimation of
depth-averaged temperature. The mean bias of heat flux in the NCEP data set could be 5–10 W/m
2
low
(negative bias) in the Middle Atlantic Bight (OAFlux group at WHOI, personal communication).
CHEN ET AL.: DIAGNOSING THE COASTAL WARMING IN 2012
6

[19] The wind stress anomalies also affect the SST
changes. The total wind speed is inversely correlated with
downward net heat flux (not shown) which is not surprising
according to (4) and (5). Furthermore, considering the shal-
low water depth at three buoys, both the along-shelf wind
and the cross-shelf wind are of dynamical importance to
the cross-shelf advection [Fewings et al., 2008 ; Horwitz,
2012], which would potentially change the advection term,
Ð0
-HurT0dz . We find that on an intraseasonal time scale,
the coupling between cross-shelf wind (positive shoreward)
anomaly and the jet stream latitude anomaly is inconsistent
at three buoys in the MAB. The correlations between the
jet stream latitude anomalies and cross-shelf wind anoma-
lies during November 2011 to June 2012 are 0.59, 0.60,
and 0.35, and decrease from Nantucket Shoals to Virginia
Beach. In particular, during March to May 2012, the cross-
shelf wind anomalies are not consistent at the three buoys.
In comparison, the correlation coefficients between the net
atmospheric heat flux anomaly and the jet stream latitude
anomaly are 0.62, 0.50, and 0.48 (at Nantucket Shoals,
Long Island, and Virginia Beach), and heat flux variations
are consistent over the entire MAB during March to May
2012. The relatively weak coupling between the cross-shelf
wind and the jet stream latitude contributes more to the net
air-sea heat flux than the cross-shelf advective flux. The
correlation between the along-shelf wind anomaly (positive
poleward) and the jet stream latitude anomaly is weak,
being 0.07, 0.21, and 0.40, from Nantucket Shoals to Vir-
ginia Beach. The orientation of the coastline varies within
the MAB, so a uniform wind direction in the MAB will
have different cross-shelf and along-shelf components. The
variation of the meridional/northward wind, versus along-
shelf or cross-shelf wind has a better correlation with the
shift of the jet stream (0.53, 0.57, and 0.46 from Nantucket
Shoals to Virginia Beach), but that does not systematically
drive onshore advection of warm slope water. Thus, wind-
driven advective flux is not responsible for the consistent
intraseasonal oscillations of SST in the spring of 2012 over
the entire region.
4. Discussion and Conclusion
[20] Using available in situ observations and reanaly-
sis data, we diagnosed the warming event over the con-
tinental shelf in the MAB in the first half of 2012. Due
to recent long-term warming trends, we chose the mean
SST of 2000–2010 for comparisons to the SST anoma-
lies in 2011–2012. The thermal anomalies would be
even larger if compared to multidecadal means. Our
presentation of the magnitude of warm conditions dur-
ing 2012 is consistent with the NOAA ecosystem advi-
sory that SST during the first half of 2012 over the
continental shelf in the MAB was the highest in 150
years of measurements.
[21] XBT profiles are only shown in the upper 50 m to
reduce the possible influence of migration of the foot of the
shelf break front [Linder and Gawarkiewicz, 1998 ; Chen
and He, 2010]. We note that the Oleander XBT data near
Jan Feb Mar Apr May Jun Jul
0
3
6
T anml [oC]
2012
44099cc1= 0.64 (6)
cc2= 0.73 (6)
−20
0
20
Lat anml [oN]
T anml
JS Lat anml
0
3
6
T anml [oC]
44025cc1=0.52 (3)
cc2=0.60 (3)
−20
0
20
Lat anml [oN]
0
3
6
T anml [oC]
44008cc1=0.56 (4)
cc2=0.84 (3)
−20
0
20
Lat anml [oN]
0
3
6
T anml [oC]
44005cc1=0.45 (3)
cc2=0.55 (3)
−20
0
20
Lat anml [oN]
Figure 5. Correlation of buoy temperature anomaly (black) and jet stream latitude anomaly (gray) dur-
ing the first half of 2012. Maximum lag correlation coefficients and lags (in days) corresponding to the
maximum coefficients are shown (in parentheses), for January to June (cc1) and March to May (cc2),
respectively.
CHEN ET AL.: DIAGNOSING THE COASTAL WARMING IN 2012
7

the New Jersey shelf break are not sampled consistently at
the 100 m isobath each month. Temporal coverage varies
from year to year as well, so temperature profiles are not
available during each of these 6 months every year. These
practical factors are unavoidable considering the sampling
process, and they may add uncertainties to the profiles.
Nevertheless, the profiles displayed in Figure 3 as well as
other profiles at shallower depths indicate the warming sig-
nal extends through the water column and the anomaly is
well represented by the overall shift of the temperature pro-
files before the development of the seasonal thermocline in
May.
[22] Closing the heat budget is difficult in this study due
to the lack of depth-dependent temperature data as well as
current measurements. The estimation of depth-averaged
temperature for 2011–2012 is largely dependent on the cli-
matological SST and depth-averaged temperature relation-
ship. This is generally valid for the seasonal scale but may
not necessarily be true for short time scales. The overlap of
Q
cum
and T 2T
0
at the Nantucket and Long Island buoys
suggests the atmospheric heat flux largely controls the tem-
perature change over monthly to seasonal scales. Accord-
ing to (1), the discrepancy between T2T
0
and Q
cum
is due
to horizontal advection. However, due to the uncertainties
in estimating depth-averaged temperature, it is difficult to
conclude that the short time scale departures between
T2T
0
and Q
cum
are entirely explained by oceanic advec-
tion. For an accurate estimate of the heat budget, full water
column data along with continuous measurements of the
currents and along-shelf and cross-shelf temperature gra-
dients are necessary.
[23] In comparison, the link between the temperature
anomaly and the jet stream related atmospheric forcing is
more definitive. The anomalous northward shift of the jet
stream during the winter of 2012 increased air temperature,
and humidity, with oscillations occurring with periods around
20–30 days. These anomalies, together with relatively weak
wind stress inhibited latent and sensible heat loss from the
ocean from their typical winter values. Presumably strong
vertical mixing during this period extended the temperature
anomalies through the water column and thus increased the
shelf-wide heat content. Strong jet stream intraseasonal oscil-
lations during March 2012 to May 2012 systematically per-
turbed the shelf temperature from the GoM to Virginia
Beach. The present analysis suggests that the ocean advective
heat flux might be secondary during this extreme event. Fur-
ther refinement of the heat budget will require focused
numerical modeling studies to account for continental shelf
and slope processes and their impact on advection of heat.
[24] The anomalous jet stream latitude in the cooling
period of late 2011 and early 2012 clearly plays an important
role in the warm anomalies. Based on the SST record at the
Nantucket buoy (44008) since the 1980s, the relationship
between the mean jet stream latitude and the mean SST dur-
ing November to February is robust (Figure 7). Besides the
anomalous 2012 condition, the relationship tends to be quasi
Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug
−200
−100
0
100
200
300
2012
heat flux [W/m2]
−10
−5
0
5
10
15
latitude
Qnet
JS
Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug
−150
−100
−50
0
50
100
150
heat flux [W/m2]
2012
latent
sensible
longw
shortw
Figure 6. (top) The jet stream latitude anomaly (gray) during 2011–2012 and net atmospheric heat flux
anomaly (red). (bottom) Different components of the air-sea heat flux, including latent heat flux (blue),
sensible heat flux (green), longwave radiation (magenta), and shortwave radiation (cyan). The positive
direction is defined as downward (into the ocean).
CHEN ET AL.: DIAGNOSING THE COASTAL WARMING IN 2012
8

linear, in that more northern jet stream positions correlate to
warmer SST near Nantucket. There is scatter in the interan-
nual variability of the relationship and the extent to which
the shift of wintertime jet stream latitude controls the sea
surface temperature variability from year to year is worthy
of future study. The quasi-linear relation shown in Figure 7
also suggests the seasonal time scale (4 month) connection
between the jet stream latitude and SST in the coastal ocean
in addition to the correlation on the intraseasonal time scale
(20–30 days) presented previously.
[25] An important aspect of the large-scale atmospheric
warming of the continental shelf is the large along-shelf
scale in the ocean. Shearman and Lentz [2010] have shown
that long-term variability of temperature anomalies is pri-
marily a two-dimensional advective process rather than a
one-dimensional process dominated by air-sea fluxes. For
the advective processes, there is a time scale of months for
signals to travel between the GoM and Cape Hatteras,
rather than the concurrent shifts from the anomalous air-sea
fluxes presented here. Further work is necessary to examine
seasonal anomalies over the past several decades.
[26] An interesting factor in the warming is the north-
ward diversion of the Gulf Stream in October 2011
[Gawarkiewicz et al., 2012; Ezer et al., 2013]. Bottom tem-
perature measurements over the continental shelf south of
New England indicate large temperature increases and thus
the heat content of the continental shelf was increased. Fur-
ther investigation of the role of this anomalous motion of
the Gulf Stream is important in order to establish the heat
content of the continental shelf before the anomalous
atmospheric warming as well as the possible influence of
anomalous SST over the continental slope in the late
autumn affecting the overlying atmospheric circulation.
[27] Based on the limited observations from 2011 to
2012, a direct link was established between the jet stream
variability and continental shelf temperature anomalies in
the MAB for the first time. Due to the importance of the
warming of ocean temperature to the marine ecosystem and
commercial fisheries, further study of the links between jet
stream variability and ocean temperature anomalies over a
longer time scale are necessary, especially within the con-
text of climate change in recent decades. Further study of
this event is important for improving management of living
marine resources during a time of rapid change in the
coastal ocean, as discussed in Mills et al. [2013].
[28]Acknowledgments. K.C. was supported by the Woods Hole
Oceanographic Institution Postdoctoral Scholarship, with funding provided
by the Cooperative Institute for North Atlantic Region. G.G.G. was sup-
ported by grant N00014-11-1-0160 from the Office of Naval Research.
S.J.L. was supported by the National Science Foundation under grant
OCE-1154575. K.C. appreciates many insightful discussions and valuable
suggestions from Young-Oh Kwon. The authors thank the NEFSC SOOP
program (http: //www.nefsc.noaa.gov/epd/ocean/MainPage/soop.html) and
the AOML High Density XBT Transect Program (http ://www.aoml.noaa.
gov/phod/hdenxbt/index.php) for the XBT data. Constructive comments
from three anonymous reviewers are much appreciated.
References
Bane, J. M., Y. H. Spitz, R. M. Letelier, and W. T. Peterson (2007), Jet
stream intraseasonal oscillations drive dominant ecosystem variations in
Oregon’s summertime coastal upwelling system, Proc. Natl. Acad. Sci.
U. S. A.,104(33), 13,262–13,267.
Barth, J. A., B. A. Menge, J. Lubchenco, F. Chan, J. M. Bane, A. R.
Kirincich, M. A. McManus, K. J. Nielsen, S. D. Pierce, and L. Washburn
(2007), Delayed upwelling alters nearshore coastal ocean ecosystems in
the northern California current, Proc. Natl. Acad. Sci. U. S. A.,104(10),
3719–3724.
Brink, K. H., R. C. Beardsley, R. Limeburner, J. D. Irish, and M. Caruso
(2009), Long-term moored array measurements of currents and hydrog-
raphy over Georges Bank: 1994–1999, Prog. Oceanogr.,82, 191–223.
Chen, K., and R. He (2010), Numerical investigation of the Middle Atlantic
Bight shelfbreak frontal circulation using a high-resolution ocean hind-
cast model, J. Phys. Oceanogr.,40, 949–964.
Emery, W. J., and R. E. Thomson (2004), Data Analysis Methods in Physi-
cal Oceanography, 654 pp., Elsevier, San Diego, Calif.
Ezer, T., L. P. Atkinson, W. B. Corlett, and J. L. Blanco (2013), Gulf
Stream’s induced sea level rise and variability along the U.S. mid-
Atlantic coast, J. Geophys. Res.,118, 685–697, doi: 10.1002/jgrc.20091.
Fairall, C. W., E. F. Bradley, D. P. Rogers, J. B. Edson, and G. S. Young
(1996), Bulk parameterization of air-sea fluxes for tropical ocean-global
atmosphere coupled-ocean atmosphere response experiment, J. Geophys.
Res.,101(C2), 3747–3764.
Fewings, M., S. J. Lentz, and J. Fredericks (2008), Observations of cross-
shelf flow driven by cross-shelf winds on the inner continental shelf, J.
Phys. Oceanogr.,38(11), 2358–2378.
Fogarty, M. J., and S. A. Murawski (1998), Large-scale disturbance and the
structure of marine systems: Fishery impacts on Georges Bank, Ecol.
Appl.,8(1), S6–S22.
Friedland, K. (2012), Ecosystem Advisory for the Northeast Shelf Large
Marine Ecosystem, Advis. 2012, No. 2., Northeast Fishery Science Cen-
ter, Woods Hole, Mass.
Gawarkiewicz, G., G. Lawson, M. Petruny-Parker, P. Fratantoni, and J.
Hare (2013), The Shelf Break Ecosystem off the Northeastern United
States: Current Issues and Recommended Research Directions, pp. 29,
Coop. Inst. for the North Atl. Reg., Woods Hole, Mass. [Available
at http://cinar.org/ fileserver.do?id5166024&pt52&p5159709.]
Gawarkiewicz, G. G., R. E. Todd, A. J. Plueddemann, M. Andres, and J. P.
Manning (2012), Direct interaction between the Gulf Stream and the shelf-
break south of New England, Sci. Rep. 2, 553, doi :10.1038/srep00553.
He, R., K. Chen, K. Fennel, G. Gawarkiewicz, and D. McGillicuddy
(2011), Seasonal and interannual variability of physical and biological
dynamics at the shelfbreak front of the Middle Atlantic Bight: Nutrient
supply mechanisms, Biogeosciences,8, 2935–2946.
Horwitz, R. (2012), The effect of stratification on wind-driven, cross-shelf
circulation and transport on the inner continental shelf, PhD thesis, 215
pp., Mass. Inst. of Technol., Cambridge.
39 40 41 42 43 44 45
6
6.5
7
7.5
8
8.5
9
9.5
10
10.5
11
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1994
1995
1996
1999
2000
2001
2002
2003
2004 2005
2006
2007
2008
2009
2012
Nov−Dec−Jan−Feb
Jet Stream Latitude
SST in the MAB [deg.C]
y = 0.37x −7.20
Figure 7. The mean jet stream latitude and the corre-
sponding mean SST during the cooling season (November
to February) at the Nantucket buoy (44008). Original data
are available from 1983. For each year, data with a gap lon-
ger than 2 months are excluded from temporal averaging.
The solid line represents the linear fit with the expression
shown on the top left.
CHEN ET AL.: DIAGNOSING THE COASTAL WARMING IN 2012
9

Ji, R., C. Davis, C. Chen, and R. Beardsley (2008), Influence of local and
external processes on the annual nitrogen cycle and primary productivity
on Georges Bank: A 3-D biological-physical modeling study, J. Mar.
Syst.,73(1–2), 31–47.
Lentz, S. J., R. C. Beardsley, J. D. Irish, J. Manning, P. C. Smith, and R. A.
Weller (2003), Temperature and salt balances on Georges Bank
February-August 1995, J. Geophys. Res.,108(C11), 8006, doi: 10.1029/
2001JC001220.
Linder, C., and G. Gawarkiewicz (1998), A climatology of the shelf-break
front in the Middle Atlantic Bight, J. Geophys. Res.,103(C9), 18,405–
18,423.
Loder, J. W., B. Petrie, and G. Gawarkiewicz (1998), The coastal ocean off
northeastern North America: A large-scale view, in The Sea, vol. 11,
edited by A. R. Robinson and K. H. Brink, pp. 105–133, Harvard Univ.
Press, Cambridge, Mass.
Lott, F., A. W. Robertson, and M. Ghil (2004a), Mountain torques and
Northern Hemisphere low-frequency variability. Part I: Hemispheric
aspects, J. Atmos. Sci.,61, 1259–1271.
Lott, F., A. W. Robertson, and M. Ghil (2004b), Mountain torques and
Northern Hemisphere low-frequency variability. Part II : Regional
aspects, J. Atmos. Sci.,61, 1272–1283.
Marra, J., R. W. Houghton, and G. Christopher (1990), Phytoplankton
growth at the shelf-break front in the Middle Atlantic Bight, J. Mar.
Res.,48(4), 851–868.
Mills, K., et al. (2013), Fisheries management in a changing climate : Les-
sons from the 2012 ocean heat wave in the Northwest Atlantic, Oceanog-
raphy,26, 191–195, doi: 10.5670/oceanog.2013.27.
O’Reilly, J. E., C. E. Evans-Zetlin, D. A. Busch, R. H. Backus, and D. W.
Bourne (1987), Primary production, in Georges Bank, pp. 220–233, MIT
Press, Cambridge, Mass.
Ryan, J. P., J. A. Yorder, and P. C. Cornillon (1999), Enhanced chlorophyll
at the shelfbreak of Mid-Atlantic Bight and Georges Bank during the
spring transition, Limnol. Oceanogr.,44(1), 1–11.
Shearman, R. K., and S. J. Lentz (2010), Long-term sea surface temperature
variability along the U.S. East Coast, J. Phys. Oceanogr.,40, 1004–
1016.
CHEN ET AL.: DIAGNOSING THE COASTAL WARMING IN 2012
10