Transport estimates of the Western Branch of the Norwegian Atlantic

advertisement
Deep-Sea Research I 79 (2013) 86–95
Contents lists available at SciVerse ScienceDirect
Deep-Sea Research I
journal homepage: www.elsevier.com/locate/dsri
Transport estimates of the Western Branch of the Norwegian Atlantic
Current from glider surveys
F. Høydalsvik a, C. Mauritzen a,n, K.A. Orvik b, J.H. LaCasce c, C.M. Lee d, J. Gobat d
a
Norwegian Meteorological Institute, P.B. 43 Blindern, N-0313 Oslo, Norway
Geophysical Institute, University of Bergen, P.B. 7803, N-5020 Bergen, Norway
c
Department of Geosciences, University of Oslo, P.B. 1047 Blindern, N-3016 Oslo, Norway
d
Ocean Physics Department, Applied Physics Laboratory, University of Washington, P.O. Box 3556400, Seattle 98105-6698, WA, USA
b
art ic l e i nf o
a b s t r a c t
Article history:
Received 2 January 2013
Received in revised form
10 May 2013
Accepted 16 May 2013
Available online 6 June 2013
The northernmost limb of the Atlantic Meridional Overturning Circulation (AMOC), so relevant for
understanding decadal climate variability, enters the Nordic Seas as the Norwegian Atlantic Current and
continues on to recirculate in the Arctic Ocean. The strength of the Eastern Branch of the Norwegian
Atlantic Current has been systematically monitored for over 15 years at the Svinøy section off southern
Norway, whereas the strength of the Western Branch has not. We therefore used autonomous gliders to
monitor and quantify the strength of this broader branch at the Svinøy section, located 500 km
downstream from the Iceland–Scotland Ridge, and at the Station Mike section 300 km further downstream. The gliders' diving depth is 1000 m, spanning the warm Atlantic Water. The current encompasses
more than warm Atlantic Water; we find that the transport peaks in two distinct temperature ranges,
one around 7.5–8 1C (Atlantic Water, carrying 7 Sv (1 106 m3/s)) and another around −0.5 1C (Norwegian Sea Deep Water, carrying 12 Sv). Contrary to earlier expectations, our results indicate that the
Western Branch carries as much water of Atlantic origin (temperature47.5 1C) as the Eastern Branch.
It should therefore be included in future monitoring plans for this region.
& 2013 Elsevier Ltd. All rights reserved.
Keywords:
Norwegian Atlantic Current
Atlantic Meridional Overturning Circulation
AMOC
Ocean monitoring
Transport
Glider
Svinøy
Ocean weather station mike
Variability
IPY
International Polar Year
1. Introduction
The Norwegian Atlantic Current (NwAC) is the northern limb of
the Gulf Stream system and carries warm and saline water of
Atlantic origin from the North Atlantic through the Nordic Seas to
the Arctic Ocean. Quantifying and understanding its variability is
important for our understanding of the regional climate system in
northern Europe and Eurasian Arctic.
The NwAC enters the Nordic Seas primarily across the Iceland–
Faroe Ridge and through the Faroe–Shetland Channel (Fig. 1). The
current continues as a two-branch system through the Nordic Seas
(Poulain et al., 1996; Orvik and Niiler, 2002). The Western Branch
of the Norwegian Atlantic Current can be considered an extension
of the Iceland–Faroe Frontal Jet which continues eastward as the
Faroe Current north of the Faroe Islands. The Faroe Current
has been monitored since 1997. The estimated average volume
transport of water warmer than 5 1C for 1997–2001 is 3.8 Sv
n
Corresponding author. Now at: CICERO Center for International Climate and
Environmental Research—Oslo, P.B. 1129 Blindern, N-0318 Oslo, Norway.
Tel.: +47 90912105.
E-mail address: c.mauritzen@cicero.oslo.no (C. Mauritzen).
0967-0637/$ - see front matter & 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.dsr.2013.05.005
(1 Sv ¼106 m3 s−1), with a peak in the temperature range 7–7.5 1C
(Hansen et al., 2003).
Current meter measurements of the Eastern Branch in the
Faroe–Shetland Channel between 1994 and 2008 yield an average
transport of 3.8 Sv (Østerhus et al., 2005), all warmer than 8 1C
(Mauritzen et al., 2011). A recent estimate of the net northward
transport between Shetland and Iceland, based on direct, repeat,
ship-of-opportunity current measurements, supports the earlier
measurements by finding that there is a net northward flow across
the section of 8.5 Sv (Rossby and Flagg, 2012).
The Eastern Branch represents a quasi-barotropic current along
the Norwegian shelf edge toward the Fram Strait, with its core
over the 500 m isobaths. The Western Branch, on the other hand,
is a baroclinic frontal jet further offshore, continuing through the
Nordic Seas toward the Fram Strait (Orvik and Niiler, 2002). There
is extensive exchange of water between the two branches (Rossby
et al., 2009), such that also the area between the branches is filled
with warm and salty Atlantic water (Fig. 1). Parts of the Iceland–
Faroe inflow actually join the Eastern Branch already in the Faroe–
Shetland Channel (Poulain et al., 1996).
Within the Nordic Seas, the warm Atlantic Water encounters
colder and fresher water masses on all sides (Fig. 1). Despite large
F. Høydalsvik et al. / Deep-Sea Research I 79 (2013) 86–95
72°N
87
67°N
66°N
68°N
65°N
64°N
64°N
63°N
60°N
62°N
2°W
0°
2°E 4°E
6°E
8°E
56°N
0
52°N
24°W
16°W
8°W
0°
5
SST (°C)
10
8°E
Fig. 1. The NwAC in the southern part of the Nordic Seas is visible in the mean SST during January–March 2009 (satellite data courtesy Steinar Eastwood, Norwegian
Meteorological Institute; Copyright (2009) EUMETSAT). The Svinøy section (extended northwestward to 11W, 65.11N) is shown, together with dots denoting the locations of
(from northwest to southeast along the section) the Seaglider offshore target (at ∼3000 m bottom depth), the 2000 m isobath, and the 1100 m isobaths (used as an inshore
limit of the Western Branch, see text). The following isobaths are shown: 500 m, 1000 m, 2000 m, and 3000 m. The zonal Station Mike section at 661N extended from the
Norwegian continental shelf, past the former Ocean Weather Station Mike to 11W (marked with a square). The 1100 m isobath is also marked with a square. The two
branches of the Norwegian Atlantic Current are sketched into the figure. The small map zooming in on our area shows the Seaglider trajectories at the Svinøy and Station
Mike sections. Key bathymetric features are labeled with acronyms: IFR—The Iceland–Faroe Ridge; FSC—The Faroe–Shetland Channel; VP—The Vøring Plateau, and VPE—
The Vøring Plateau Escarpment.
temporal variability in temperature and salinity within the key
water masses in this region during the 20th century (Dickson and
Østerhus, 2007), a temperature–salinity diagram reveals a welldefined transition between the warm and saline waters of Atlantic
origin and the colder and fresher surrounding water masses in the
vicinity of S ¼35 (Fig. 2). Therefore there exists a strong tradition,
stemming from Helland-Hansen and Nansen's seminal work
“The Norwegian Sea” (1909), to define Atlantic Water in the
Norwegian Sea as water with salinities higher than 35 (Fig. 2),
corresponding to temperatures higher than 4–5 1C in the southern
Nordic Seas and colder further north. On the eastern side, the
shallow Norwegian Coastal Current runs northward along the
coast of Norway from the Baltic, picking up river runoff along
the way (Mork, 1981). Its salinity is typically less than 34.8, and in
the winter its temperature is in the 2–5 1C range (Saetre and
Ljoen, 1972). On the western side, colder waters of Arctic origin
enter the Nordic Seas in the western Fram Strait as Polar Water
with salinities lower than 34.5 and temperatures around 0 1C. All
these water masses are found in the upper ocean, and there are
broad regions of the upper Nordic Seas that consist of mixtures of
these water masses. The deep waters of the Nordic Seas originate
in the Greenland Sea and in the Arctic Ocean. Salinities are
typically around 34.9 and temperatures less than 0 1C (Aagaard
et al., 1985) (Fig. 2).
The standard “Svinøy section” (Fig. 1) captures the Norwegian
Atlantic Current about 500 km downstream from the Iceland–
Scotland Ridge. The temperature and salinity of this section has
been observed several times a year for more than 50 years, as part
of the Norwegian Institute of Marine Research's standard hydrographic monitoring program (www.imr.no; see also Mork and
Blindheim, 2000). In addition, the current strength of the Eastern
Branch of the NwAC has been monitored continuously since 1995
with moored current meters. The average transport estimate for
water warmer than 5 1C is 4.4 Sv and the transport peaks in the
temperature range 8.5–9 1C (Orvik et al., 2001; Orvik and Skagseth,
2005; Mauritzen et al., 2011). The temperature range of the
Eastern Branch at the Svinøy section is within the temperature
Norwegian Coastal Water
Atlantic Water
Polar Water
Deep Water
Fig. 2. Potential Temperature–Salinity diagram, showing the main water masses of
the Nordic Seas. Based on the World Ocean Atlas 2005 (Antonov et al., 2006;
Locarnini et al., 2006). Also shown are lines of constant sθ [kg/m3].
range of the two inflow branches at the Iceland–Scotland Ridge,
but with about half of the total warm water volume transport at
the ridge. Achievements of accurate estimates of the AI to the
Nordic Seas have been addressed in a series of papers over the past
decades using different methodologies as budget considerations
(e.g. Mauritzen (1996); Worthington, 1970) and direct measurements over the ISR (Hansen and Østerhus, 2000) and in the Svinøy
section just to the north of the FSC (Mork and Blindheim, 2000; Orvik
et al., 2001); converging toward an overall estimate of about 8 Sv.
Transport estimates of the Western Branch at the Svinøy
section are sparse, and mainly based on dynamic calculations
88
F. Høydalsvik et al. / Deep-Sea Research I 79 (2013) 86–95
from the hydrography monitoring progam. Applying a level of no
motion for the deep layer, and considering waters warmer than
1 1C, Mork and Blindheim (2000) obtained 2.7 Sv for the period
1955–1996. Orvik et al. (2001), using a different method but the
same deep level of no motion, obtained 3.4 Sv (for water warmer
than 5 1C). Orvik (2004) demonstrated that using a level of no
motion may lead to an underestimate of the transport, since the
baroclinic flow rides on a cyclonic circulation of order 0.1 ms−1
along closed isobaths in the Norwegian Sea. Geostrophic estimates
of total Norwegian Atlantic Current transport thus require accurate
estimates of reference level velocity. Recently, Mork and Skagseth
(2010) combined absolute dynamic topography (from satellite
altimetry) and hydrography and produced an updated, remarkably
low (1.7 Sv), estimate of the transport of the Western Branch for
waters more saline than 35. A salinity of 35 is in this region
typically comparable to a temperature of 5 1C, the number used by
Orvik et al. (2001). Thus the range of the estimates of the strength
of the Western Branch is very large.
Underwater gliders provide an alternative approach to estimate
absolute geostrophic velocity, yielding measurements both of the
vertically averaged velocity during the dive as well as of the
geostrophic shear (Todd et al., 2011). The gliders (see e.g. Rudnick
et al. (2004)) are autonomous underwater vehicles that change
their buoyancy by inflating and deflating a bladder, while using lift
generated by the wings and body to convert the resulting vertical
motion into horizontal motion. When at the surface, the gliders
geolocate using GPS and exchange data and new commands with a
base station located on shore using an Iridium satellite telephone.
The gliders steer by adjusting their attitude (pitch and roll),
navigating between waypoints while profiling up and down in a
saw-tooth pattern. Gliders can operate for several months by
moving slowly and by carefully managing power. They have been
used in a wide range of studies, including eddy processes (see e.g.
Martin et al. (2009)), ecosystems (see e.g. Alkire et al. (2012),
Briggs et al. (2011)), and hydrographic monitoring (see e.g. Perry
et al. (2008)).
During the International Polar Year (2007–2009), we used
gliders to monitor the Norwegian Atlantic Current in the Nordic
Seas for the first time. During a period of nearly a year (2008–
2009), we obtained nine glider transects at the Svinøy section in
the southeastern Norwegian Sea and three transects along 661N,
the latitude of the historical Ocean Weather Station Mike (661N,
21E; see Dinsmore (1996)) (Fig. 1). Taking advantage of modern
glider technology and analysis methods, we calculate the horizontal and vertical distribution of temperature, salinity and
velocity as well as volume transport at these two crossings of
the Norwegian Atlantic Current.
The paper is organized as follows: In Section 2 we describe the
glider operations and instrumentation, as well as the methods of
analysis. In Section 3 we describe the results of the analysis, in
terms of the horizontal and vertical structure of the current.
We isolate the transport of Atlantic Water using a thermal space
analysis, and discuss implications for entrainment. We then
investigate the time variability of the Atlantic Water transport.
Finally, in Section 4 we discuss the results.
2. Dataset and methods
2.1. Gliders
In this study we used the seaglider, an underwater glider that
has been under continual development at the University of
Washington since the mid-1990s (e.g. Eriksen et al. (2001)).
The seagliders were instrumented with a Seabird CTD. Typical
Table 1
Details of the Station Mike (TM1–TM3) and Svinøy (T1–T9) transects: time period
for each transect, dive numbers. Transects with odd numbers are onshore transects,
while transects with even numbers are offshore transects.
Transect
Period
Transect dives
All dives
TM1
TM2
TM3
T1
T2
T3
T4
T5
T6
T7
T8
T9
03.08.08–24.08.08
31.08.08–15.09.08
16.09.08–05.10.08
10.02.09–02.03.09
12.03.09–28.03.09
28.03.09–12.04.09
19.04.09–03.05.09
03.05.09–19.05.09
27.05.09–09.06.09
09.06.09–30.06.09
08.07.09–22.07.09
22.07.09–07.08.09
145–213
245–293
294–354
57–118
156–206
207–256
281–324
325–376
402–443
444–512
540–584
585–635
145–223
224–293
294–370
57–137
138–206
207–273
274–324
325–378
379–443
444–526
527–584
585–644
dives lasted 7–8 h, ranging to 1000 m depth, while covering a
horizontal distance of 3–7 km, depending on the ambient currents.
Seaglider 17 was deployed on July 4, 2008 and operated the
zonal “Station Mike section” along 661N. This extends from the
Norwegian continental shelf, past Ocean Weather Station Mike, to
11W (Fig. 1). Three transects were completed before the glider was
recovered on October 5. Seaglider 160 was deployed on January 24,
2009 at the Svinøy section, and completed nine transects before
recovery on August 10. All dives extended to a depth of roughly
1000 m, well below the depth of Atlantic Water at these locations.
Table 1 summarizes the details of transects and dives.
The glider surveys initially targeted both the Eastern and
Western branch of the Norwegian Atlantic Current, but strong
currents (depth-averaged velocities larger than 0.4 m s−1) in the
Eastern Branch prevented the glider from staying on track there.
However, as mentioned earlier, the Eastern Branch of the NwAC
has been monitored continuously since 1995 by moored current
meters at Svinøy section. This analysis therefore focuses on the
Western Branch of the NwAC.
2.2. Velocity and transport calculations
The glider employed a depth-dependent sampling scheme that
collected measurements at intervals ranging from finer than 0.5 m
at depths shallower than 60 m to 4 m intervals below 500 m.
Measurements from ascending and descending profiles were
averaged in 5 m vertical bins to reduce noise and facilitate
analysis. For each pair of dives, we averaged measurements from
the ascending profile of one dive and the descending profile of the
next to produce the binned vertical profile. This profile was
assigned to the middle position between the end-coordinates of
first dive and the start-coordinates of the next dive. To minimize
the noise associated with ageostrophic effects, e.g. the effect of
tidal displacements, the profiles and depth-averaged currents
were smoothed by using a ten-dive running mean filter, yielding
a horizontal resolution of 30–70 km.
The cross-track baroclinic velocities are calculated from the
density measurements, and the reference level velocity is determined by matching the cross-track surface displacement of the
glider (see Appendix A). The geostrophic transport per unit track
length is then the vertical integral of the velocity, and the total
cross-track transport is obtained by integrating along the track
from the start to end of each section. The dives and endpoints used
in the transport calculations for the Svinøy and Station Mike
sections are shown in Table 1 and depicted in Fig. 3. As explained
in Appendix A, we use here the velocity component at the
maximum diving depth as a proxy for the “barotropic” velocity.
F. Høydalsvik et al. / Deep-Sea Research I 79 (2013) 86–95
67°N
67°N
100 km
66°N
66°N
66°N
2°W
0°
2°E
67°N
65°N
2°W
6°E
4°E
0.3m/s
0°
30’
2°W
66°N
66°N
30’
30’
30’
2°W
0.3m/s
65°N
0°
2°E
67°N
4°E
2°W
6°E
30’
0.3m/s
0°
2°E
4°E
2°W
6°E
66°N
66°N
30’
30’
30’
2°W
0.3m/s
65°N
0°
2°E
67°N
4°E
2°W
6°E
0°
2°E
4°E
66°N
30’
30’
30’
65°N 0.3m/s
2°E
4°E
6°E
2°W
2°E
4°E
6°E
100 km
30’
66°N
0°
0°
67°N
100 km
66°N
2°W
6°E
100 km
2°W
6°E
30’
65°N 0.3m/s
4°E
0.3m/s
65°N
67°N
100 km
30’
2°E
30’
30’
0.3m/s
0°
67°N
100 km
66°N
65°N
6°E
4°E
100 km
65°N
67°N
100 km
2°E
30’
66°N
0.3m/s
0°
67°N
100 km
30’
65°N
0.3m/s
65°N
6°E
4°E
2°E
67°N
100 km
1000
65°N
30’
30’
2000
0.3m/s
1000
2000
30’
100 km
30’
30’
1000
100 km
30’
2000
67°N
89
65°N
0°
2°E
4°E
6°E
0.3m/s
2°W
0°
2°E
4°E
6°E
Fig. 3. Seaglider depth-averaged velocities [m/s] from transect 1–3 at the Station Mike section (a–c), and transects 1–9 at the Svinøy section (d–l), shown as quiver plots.
The parts of the transect that were included in the analysis are inside the yellow brackets. The isobaths range from 500 m to 3500 m, with intervals of 500 m. (For
interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
The three transects at the Station Mike section, and in particular the nine transects at the Svinøy section, stayed relatively
close to the target track (the direct line between the onshore and
offshore targets) (Fig. 3). In order to estimate the average state of
the current (for instance in Fig. 4) we therefore project the
observations from each transect onto the target track and then
linearly interpolate to regular 500 m horizontal intervals, before
finally averaging the projected transects.
2.3. Error estimation
Uncertainties associated with glider-based depth-average currents represent the largest source of error in the estimated
transports. The difference between the observed displacement,
calculated from GPS positions taken at the start and end of each
dive, and the displacement estimated from a hydrodynamic model
that simulates glider motion from vehicle buoyancy, pitch, roll and
heading, provides an estimate of velocity averaged over the profile
(Eriksen et al., 2001). The error in the modeled displacement,
which depends on the validity of the hydrodynamic model and the
fidelity of the inputs, leads to uncertainty in the depth averaged
current. Eriksen et al. (2001) suggest typical uncertainties of
1–1.5 cms−1. Though there are no a priori reasons to expect these
errors not to cancel out, we estimate, conservatively, that that they
do not cancel out and therefore correspond to a net transport
uncertainty of 1.1 Sv for the 76 km2 Svinøy section.
We found, in our data, consistent differences between transport
estimates calculated from onshore and offshore transects, indicating a compass heading bias. Hard iron effects (those that carry
their own magnetic field) are heading dependent and typically
dominate glider compass errors. Although heading-dependent
errors complicate the task of developing corrections, the problem
was simplified in the Svinøy section missions because the glider
repeatedly occupied a single line between a pair of reciprocal
headings and because it remained very close to the direct line
(Fig. 3). In this case, the problem is analogous to correcting for
transducer misalignment relative to the hull in ship-based Acoustic Doppler Profiler measurements (Joyce, 1989). A single compass
F. Høydalsvik et al. / Deep-Sea Research I 79 (2013) 86–95
600
800
150
200
400
35.1
35
35.2
600
35.1
800
35
1000
250
34.9
50
100
s (km)
Depth (m)
Depth (m)
0
10
400
600
100
150
200
1000
250
100
100
150
200
27.8
400
250
Velocity (m/s)
Velocity (m/s)
27.6
0.2
0.1
0
27.8
28
0
0.1
−0.1
−0.2
50
100
150
200
250
200
250
200
250
0.15
0.1
0.05
0
−0.05
50
100
150
200
250
50
100
Bottom depth (m)
1000
1500
2000
2500
3000
1000
1500
2000
2500
3000
3500
50
100
150
150
s (km)
s (km)
Bottom depth (m)
26.4
26.6
26.8
27
27.2
27.4
27.6
26.4
26.6
26.8
27
27.2
27.4
0.2
vda
v800
v100
0
3500
250
s (km)
0.05
−0.05
200
28
800
0.2
0.1
150
600
s (km)
0.15
2
27.8
28
1000
50
26.6
26.8
27
5 27.4
27.2
0.0 27.6
0.1
Depth (m)
0.05
1000
28
0.05
Depth (m)
0.05
600
200
27.6
27.8
28
800
4
0.1
28
27
27.2
27.4
27.4
27.6
27.8
0.15
400
0.1
5
0.1
27.8
6
s (km)
0.05
27.4
27.6
8
4
2
0
50
s (km)
200
6
0
800
50
12
11
8
800
1000
250
200
400
600
200
s (km)
8
6
4
2
200
150
Temperature (°C)
100
35.3
Velocity (m s−1)
50
35.3
35.2
0.0
5
1000
200
0.05
Depth (m)
400
Depth (m)
35.
3
3
35.15.2
35
200
Salinity
90
200
250
s (km)
50
100
150
s (km)
Fig. 4. Average fields from the Svinøy transects (a–e) and Station Mike transects (f–j), calculated by projecting each transect onto the target track and then averaging (see
Section 2.2). (a) and (f): salinity; (b) and (g): potential temperature [1C]; (c) and (h): cross-track absolute geostrophic velocity (colored and contoured) [m/s] and along-track
potential density (sθ) (contoured) [kg/m3]; (d) and (i): Cross-track absolute geostrophic velocity at 100 and 800 m depth as well as the depth-averaged velocity [m/s]; (e) and
(j): mean bottom depth [m] obtained from averaging the transect along-track bottom depths.
bias, applicable to the pair of headings, can be estimated by
minimizing some measure of change between measurements
collected while moving along the two different headings (Todd
et al., 2011). Following the approach used for correcting shipmounted ADCPs (Joyce, 1989), Todd et al. (2011) chose to minimize
differences in depth-average velocity in the profiles just before
and after the turns. We could not use this approach at the inshore
turns because of the strong currents of the Eastern Branch in that
vicinity, leaving us only with the offshore turns, and that failed to
provide enough data.
As an alternative, we chose to minimize the difference in
transect-integrated transport between successive complete occupations. This provided a relatively stable estimate of compass bias,
albeit with the likely tradeoff of damping some of the true
temporal variability. Corrections were derived by computing mean
cross-track, depth average velocity for a range of heading biases,
using different biases for dives (pitch down) and climbs (pitch up)
to allow for attitude-dependent compass errors (e.g., uncorrected
soft iron effects; those that distort the Earth's magnetic field, but
do not generate a magnetic field of their own). The corrections
F. Høydalsvik et al. / Deep-Sea Research I 79 (2013) 86–95
91
Table 2
Volume transports (Sv) through the Svinøy section offshore of the 1100 m isobaths (see text for details). Atlantic Water is defined as having salinity higher than 35.
Total transport
Atlantic Water (S435) transport
Barotropic component of Atlantic Water transport
T1
T2
T3
T4
T5
T6
T7
T8
T9
Average
21.1
6.8
4.5
23.8
8.4
5.7
17.7
6.4
3.7
19.7
7.1
3.6
19.6
7.7
3.5
18.2
7.3
2.9
20.9
7.4
4.4
14.2
5
1.9
17.8
5.2
2.6
19.2 7 2.7
6.8 7 0.37
3.6 7 0.38
28.5
28
1
27
0.5
26.5
0
σ0 (kg m−3)
Volumeflux (Sv)
27.5
13
12
34.5
34.6
34.7
34.8
34.9
35
35.1
35.2
35.3
35.4
35.5
11
10
9
8
7
Sa
6
lin
5
ity
4
3
2
1
0
−1
P
n
ote
tia
m
l te
pe
r
26
re
atu
25.5
25
Fig. 5. Total volume transport [Sv] at the Svinøy section, calculated as a function of temperature and salinity (intervals of 0.5 1C, 0.05 salinity units, respectively). Transports
are calculated down to 1000 m. The potential density (sθ) is indicated in color. The numerical values are given in Table 2.
that minimized the Root Mean Square difference between sequential mean cross-track, depth-average velocity were 4.01 when
pitched up and 5.251 degrees when pitched down. These corrections were applied to the heading record and derived quantities
(e.g. depth average velocity) recomputed.
Other potential sources of uncertainty are smaller; more details
are given in Appendix B. The error bars used in Table 2 are
standard errors of the mean (standard deviations divided by the
square root of the number of observations).
3. Results
The depth-averaged glider velocities for the three Station Mike
transects and for the nine Svinøy transects are shown in Fig. 3 (a–c,
d–l, respectively). Through all transects there is a net northward
velocity, reflecting the general direction of the Norwegian Atlantic
Current. The glider in some cases captures not only the Western
Branch but also the offshore part of the Eastern Branch of the
NwAC (Fig. 3b, d and e).
At the Svinøy section the large temporal and spatial variability
of the current is evident, ranging from a distinct jet (Fig. 3d) to a
current which is nearly uniform laterally (Fig. 3e and l). Yet, the
offshore part of the Eastern Branch at the Svinøy section (Fig. 3d
and e) is always found inshore of the 1000 m isobath, and we find
that there is typically a current minimum in the vicinity of the
1000 m isobaths, consistent with results by Orvik et al. (2001) who
found that the Eastern Branch has a well-defined offshore limit
around 1000 m. We therefore define the inshore edge of the
Western Branch as lying over the 1100 m isobath (see Sections
3.2 and 3.3; see also Appendix B3).
3.1. Spatial structure of the Western Branch
The horizontal and vertical structure of the hydrography and
velocity cores of the Western Branch at the Svinøy and Station
Mike sections are most clearly seen in the two time-mean glider
sections (Fig. 4). At both locations, warm water is found in the
surface layer much farther west than the velocity core (compare
Fig. 4b and g to c and h). The velocity core at the Svinøy section is
roughly 50 km wide and 400 m deep (Fig. 4c and d), centered
between the 1500 and 2000 m isobaths, with a maximum speed of
roughly 0.2 m/s. This velocity core is associated with steep
isopycnals (Fig. 4c), demonstrating its baroclinic structure. Here,
under the most pronounced surface currents, we find the weakest
deepwater velocities (beneath 600 m; Fig. 4c). Shoreward and
seaward of the frontal region the flow has less vertical shear
(appears more barotropic).
At the Station Mike section the velocity core is located offshore
of the steepest topography, between the 2000–3000 m isobaths
(Fig. 4h–j). The core appears more barotropic here than at the
Svinøy section. The maximum average deep velocity at 800 m is
92
F. Høydalsvik et al. / Deep-Sea Research I 79 (2013) 86–95
0.12 ms−1, roughly twice the corresponding deep velocity of the
Svinøy section (contrast Fig. 4d and i). Although this difference
could be attributed to noise stemming from the relatively small
number of realizations averaged to create the Station Mike section,
this northward intensification of the deeper flow is consistent
with the results of Søiland et al. (2008), who found an intensification in the velocity of RAFOS floats at 800 m as they moved from
the area of the north-western part of the Svinøy section and into
the area of the Vøring Plateau Escarpment (VPE, Fig. 1).
and Nansen's (1909) classical definition of Atlantic Water (S4 35)
to separate the two transport peaks.
It is Norwegian Sea Deep Water that dominates the Western Branch of the Norwegian Atlantic Current at the Svinøy
section: According to our dataset the deep cyclonic circulation of
the Norwegian Sea Deep Water adds about 12 Sv to the total
transport through the section (Table 2; Fig. 5). The mean transport
of Atlantic Water in the Western Branch, as measured by the gliders,
is 6.8 Sv (Table 2, Fig. 5). Dividing the Atlantic Water transport into its
“baroclinic” and “barotropic” components according to the definition
Appendix A we find that the “barotropic” component contributes
more than 50% to the transport (3.6 Sv, see Table 2), and that the
“baroclinic” component contributes 3.2 Sv.
3.2. Transport of the Western Branch of the Norwegian Atlantic
Current at the Svinøy section
The current we have described encompasses more than just
warm Atlantic Water. Comparing the temperature and velocity
fields in Fig. 4 reveals that there is not a one-to-one match
between high velocities and high temperatures. The total average
transports observed by the gliders at the Svinøy and Station Mike
sections are 19 Sv (Table 2) and 11 Sv, respectively, numbers that
are highly dependent upon how far west into the gyres the
offshore endpoins of the glider sections were set. More important,
in the context of the Meridional Overturning Circulation, is to
estimate the transport of Atlantic Water. We will focus specifically
on the Svinøy section, since we have many more transects there
than at the Station Mike section.
To isolate the Atlantic Water component of the NwAC, we
estimate the average state of the current as a function of
temperature. This approach is instructive because of the large
temperature contrasts between waters of Atlantic and Arctic
origins in the region. We calculate the transport within temperature bins of 0.5 1C and salinity bins of 0.05 (see Fig. 5), and find
that in temperature space the transport is distributed into two
distinct peaks. One transport peak is the warm and relatively light
Atlantic Water, with temperatures in the range 7.5–8.0 1C. The
other peak is the dense Norwegian Sea Deep Water (HellandHansen and Nansen, 1909), with temperatures near −0.5 1C and
salinities near 34.92. There is very little transport of waters with
salinity near 35, which here is in the temperature range 3.5–4.5 1C
(Fig. 5). We therefore find that we can still use Helland-Hansen
3.3. Correlation with local winds
The Norwegian Atlantic Current exhibits short-term temporal
variability. For instance the transport in the Eastern Branch
increases by roughly 20% in winter (Jakobsen et al., 2003; Orvik
and Skagseth, 2005; Andersson et al., 2011). Variations in the
NwAC have been linked previously to remote wind forcing over
the North Atlantic (Orvik and Skagseth, 2003; Skagseth et al.,
2004; Olsen et al., 2008; Richter et al., 2009). Likewise, variability
in the currents in the closed gyres of the Nordic Seas (the
Norwegian, Lofoten and Greenland Basins) has been found to
correlate with local wind forcing within the Nordic Seas (Isachsen
et al., 2003; Furevik and Nilsen, 2005). Less is known about the
relationship between the NwAC and the local wind forcing in the
Nordic Seas.
We test that here, by comparing the observed “barotropic”
transports with the wind stress curl, integrated over the southeastern portion of the Nordic Seas. For the winds, we use the
ERA40 ten-meter values from ECMWF in combination with a bulk
formula for the wind stress, with a drag coefficient as specified by
Trenberth et al. (1990). We then take the curl and integrate over
the Norwegian and Lofoten basins, north of the Iceland–Scotland
Ridge and west of the Norwegian shelf break (the results do not
depend sensitively on the exact choice). The resulting integrated
wind stress curl (IWSC) time series was then smoothed with a
running mean (boxcar) filter, to remove high frequency variations.
x 104
IWSC
Barotropic transport
10
6
4
Transport (Sv)
5
Transport (Sv)
IWSC (kg ms−2)
6
4
0
2
2
0
0
−4
Barotropic transport
Linear fit
−5
Jan09
Apr09
Jul09
Oct09
−2
0
2
IWSC (kg ms−2)
4
6
8
x 104
Fig. 6. The barotropic component of the Atlantic Water (S 435) volume transport [Sv] for the nine Svinøy transects as well as the Integrated Wind Stress Curl (IWSC) [kg m/
s2] over the Nordic Seas, filtered with a 30 days boxcar filter, plotted vs. time (Left panel). The volume transport for each transect has been assigned to the transect's mean
point in time. The barotropic Atlantic Water (S435) volume transport plotted versus the IWSC. A linear fit is also shown (Right panel).
F. Høydalsvik et al. / Deep-Sea Research I 79 (2013) 86–95
93
Table 3
Volume transport [Sv] at the Iceland–Scotland Ridge and in the Svinøy section as in Mauritzen et al. (2011), but here with a minor update for the Svinøy section West after
the correction of the compass direction and applying a ten-dive tidal filter.
Volume transport
45 1C
46.5 1C
47 1C
47.5 1C
48 1C
Temperature range of transport peak
Period of monitoring
FSC
IFR
Svinøy East
Svinøy West
4
3.8
4.4
5.8
4
2.8
3.7
4.4
4
2.35
3.3
3.8
4
1.7
2.8
2.8
4
1.3
2.1
1.6
9.5–10
7–7.5
8.5–9
7–8
1994–2008
1997–2001
1995–1999
2009
The result is plotted with the observed transports in the
left panel of Fig. 6. The IWSC has been scaled for comparison.
The transport time series is short, allowing for only a limited
period for comparison, but during that time period the two time
series are similar. The IWSC peaks in late March and again in
September, and has a pronounced minimum in July. The observed
transport likewise exhibits its largest value in March and its
smallest value in July. There is an exception in June when the
“barotropic” transport is large and the IWSC is relatively small.
However, plotting the transport against the IWSC at the corresponding times (Fig. 6, right panel) reveals that the two time series
are correlated, with a correlation coefficient r ∼0.7. If we had
neglected the outlying point in June, the correlation would be even
higher. Thus these results indicate that the monthly “barotropic”
variations in the NwAC may be wind-driven. However, the present
result should be viewed as suggestive, due to the shortness of the
time series.
4. Discussion and conclusions
This work was motivated by a need to determine transport
values of the entire Norwegian Atlantic Current at the Svinøy
section. Using the classical definition of Atlantic Water (S 435)
and adding the Western Branch transport estimate (6.8 Sv;
Table 2) to the already published transport numbers of the Eastern
Branch (4.4 Sv; Orvik et al., 2001) yields a total Atlantic Water
transport of 11.2 Sv (Table 3). This is almost 3 Sv more than the
combined transport for the two branches at the Iceland–Scotland
Ridge (8–8.5 Sv; Table 3; Mauritzen et al., 2011; Rossby and Flagg,
2012). This in turn implies a substantial entrainment of water en
route from the ridge to the Svinøy section (Table 3), a result
consistent with Oliver and Heywood (2003) who demonstrated
that due to mixing and entrainment the through-flow of water
with salinity above 35.0 is significantly larger than the inflow of
water of North Atlantic origin. Based on the glider data we
estimate that the recirculation of colder water masses in the
Nordic Seas contributes roughly 30% (3 Sv vs. 11.2 Sv) to the
transport of water with salinity higher than 35.0 at the Svinøy
section.
If we instead define Atlantic Water by its characteristics at the
Iceland–Scotland Ridge, i.e. waters with temperature higher than
7.5 1C, the transport of Atlantic Water in both the Western and
Eastern branches at Svinøy section are reduced to 2.8 Sv (Table 3).
This is sensible, as the total transport is comparable to that at the
Iceland–Scotland Ridge. It also implies that there is little atmospheric cooling occurring during the short transit from the Ridge
to the Svinøy section. Using this definition (T 47.5 1C), both
branches are actually equally important for the transport of
volume and heat through the Nordic Seas. The transport difference
is largest for the Western Branch (6.8 Sv at S4 35.0 and 3 Sv at
T 47.5 1C), indicating that the entrainment of colder water is
largest into that branch.
Returning to the classical definition of Atlantic Water (salinity 435; Helland-Hansen and Nansen, 1909), for which the
transport is found to be 6.8 Sv (Table 2), we found in Section 3.2
that the “barotropic” component of the Western Branch Atlantic
Water transport accounts for more than 50% of the average
transport (Table 2). If we had assumed that there were a level of
no motion at 1000 m and referenced the geostrophic velocities
accordingly, our transport estimate would have been not 6.8 Sv
but 3.2 Sv (Section 3.2; Table 2). This number is close to the 3.4 Sv
found by Orvik et al. (2001), who also used a deep level of no
motion. We therefore agree with the Orvik (2004) statement that
3.4 Sv must be an underestimate of the strength of the Western
Branch since we do not find there to be a deep level of no motion.
We deem it unlikely that the Western Branch can be as weak as
1.7 Sv as suggested by Mork and Skagseth (2010) (who also used
the salinity 435 definition).
We find that the “barotropic” component of the Atlantic Water
(S 435) transport in the Western Branch covaries with the
integrated wind stress curl in the southeastern Nordic seas.
This is in line with previous studies, which found that the flow
in the closed gyres (like the Norwegian and Lofoten basins) is
forced by the integrated wind stress curl (Isachsen et al., 2003).
What is interesting is that the flow here is not confined to the
gyres, but straddles f/H contours which extend all the way into the
Arctic. So the winds may be producing similar effects over these
contours as well. However, longer time series are required to
establish this result. Note too that we have not addressed variations in the baroclinic portion of the flow, which also would
require longer time series.
Our results indicate that previously published values for the
transport within the Western Branch of the Norwegian Atlantic
Current should be revised upward to reflect the significant contribution of deep currents. Our assessment of the water mass
characteristics in the Western Branch leads us to conclude that
this branch of the Norwegian Atlantic Current carries as much
warm water as the Eastern Branch. The results presented here
nonetheless rely on the analysis of relatively sparse data, only
covering six months from one particular year. It would therefore
be very useful to include the Western Branch in any future
monitoring plan for this region.
Acknowledgments
This investigation was a part of the project entitled Integrated
Arctic Ocean Observing System Norway (iAOOS-Norway), funded
by the Norwegian Research Council. Dr. A. K. Sperrevik is acknowledged for her help with the tidal data. This research benefited
from scientific discussions with Drs. P.E. Isachsen, J.E. Weber, Ø.
Godøy, Ø. Sætra, Inga Koszalka, and O.A. Nøst. Engineers and
Seaglider pilots at APL-UW, G. Shilling, A. Huxtable, A. Wood, and
K. van Thiel have participated in navigation, technical support,
maintenance, and field work, all crucial to our research. The
Norwegian Coast Guard has assisted in both deployments and
recoveries. The Norwegian Coastal Administration has provided
storage facility for glider equipment. These two entities are greatly
acknowledged for their professional help, making this research
possible.
94
F. Høydalsvik et al. / Deep-Sea Research I 79 (2013) 86–95
Appendix A. Velocity and transport calculations
B.2. Tides
Starting from the cross-track component of the thermal wind
equation, applying the Boussinesq approximation, we obtain
Using the Oregon State University Tidal Prediction Software
(version TPXO7.1), we simulate the barotropic tidal current for the
tidal constituents M2, S2, N2, K2, O1, P1, and Q1, for each position
and time given from the Seaglider data set. These predictions are
based on the barotropic inverse tidal solutions obtained with a
Tidal Inversion Software (Egbert and Erofeeva, 2002). The obtained
values are then interpolated linearly to a grid that is regular in
time, in order to estimate the dive-by-dive average tidal current.
The depth-independent tidal currents over a few dives are found
to be of order 1 cm/s and thus negligible compared to the depthaveraged currents.
To minimize the ageostrophic noise, in particular the effect of
the internal tide heaving, we apply a ten-dive (or ten-profile)
running average filter, roughly corresponding to a three-day time
filter, to the temperature, salinity and depth-averaged currents
prior to the calculations. The transport estimates with and without
pre-filtering differ with a root mean square of 0.4 Sv.
ρ0 f
∂vn
∂ρ
¼ −g
∂s
∂z
ð1Þ
where s is the along-track coordinate (positive in the direction of
the glider path) and z is the vertical coordinate (positive upward).
vn(z) is the cross-track velocity, f is the Coriolis parameter, g is the
acceleration of gravity, ρ is the density and ρ0 is a reference
density. By integrating Eq. (1) from the maximum diving depth H
to the depth z, we obtain
Z z
g
∂ρ
dz
ð2Þ
vn ðzÞ ¼ vn ð−HÞ−
ρ0 f −H ∂s
We now assume that the vertical integral of vn(z) over the
depth of the dive H equals the depth-averaged velocity V derived
from the movement of the glider:
V¼
1
H
Z
0
−H
vn ðzÞ dz
ð3Þ
By integrating Eq. (2) over the water column, and utilizing Eq.
(3), we obtain the velocity at maximum diving depth,
vn ð−HÞ ¼ V þ
g
ρ0 f H
Z
0
−H
Z
−H z^
∂ρ
dz^ dz:
∂s
ð4Þ
The geostrophic transport per unit along-track length is
obtained from vertical integration over the water column, and
the total geostrophic cross-track transport is obtained by integrating laterally along the track from start to end of each section. The
dives and endpoints used in the transport calculations for the
Svinøy and Station Mike sections are shown in Table 1 and
depicted in Fig. 3. Although formally the current and its volume
transport are baroclinic, i.e. density surfaces and pressure surfaces
intersect, we will refer to the velocity component at maximum
diving depth, vn(–H) as the “barotropic velocity component”, and
the one associated with shear as the “baroclinic velocity
component”.
Appendix B. Error analysis associated with the transport
estimates
B.1. Ekman drift
We have examined wind data from three weather stations in
the vicinity of the glider transects: Ocean Weather Station Mike
(661N, 21E), Heidrun (65.31N, 7.31E), and Draugen (64.31N, 7.81E)
from 1999 to 2009. The six-hourly observations were interpolated
linearly to an hourly grid and averaged with a 13 h running
average filter—a period close to the inertial period of approximately 13.1 h in this area. The Ekman transport estimates were
based on the same bulk formula for the surface wind stress as in
Section 3.3. The mean Ekman current that the glider experiences
during a dive equals the Ekman transport divided by the diving
depth (this is true here because the dive depth is always larger
than Ekman penetration depth in this area). Typically, the dive-bydive effect of the Ekman drift is small, with current components of
order 1 cm/s. The maximum effect of the Ekman current on
transect time scales constitutes transports that are within the
transport standard errors, also during periods with intense winter
storms.
B.3. Other potentially important error sources and uncertainties
The errors associated with binning of the data (Section 2.2) are
small, due to the very high data sampling resolution, both along
the track and in the vertical. The error associated with ignoring the
surface drift between the dives is relatively unimportant for
transects with deep dives as in our case. From dive and surface
drift lengths we estimate this error to be about 6% or less, yielding
an error in the mean transport less than 0.4 Sv.
Not all the transects reached the 1000 m isobath in the Svinøy
section, defined as the offshore limit of the Eastern Branch (see
Section 2.2). We therefore use the 1100 m isobath when defining
the inshore limit in the transport calculation (Sections 2 and 3).
The error associated with not covering the 1000–1100 m isobath
range is estimated by utilizing data from the transects that did
reach the 1000 m isobath. The root mean square error range is
0.6 Sv.
References
Aagaard, K., Swift, J.H., Carmack, E.C., 1985. Thermohaline circulation in the Arctic
Mediterranean Seas. J. Geophys. Res. 90 (C3), 4833–4846.
Alkire, M.B., D'Asaro, E., Lee, C.M., Perry, M.J., Gray, A., Cetinic,́ I., Briggs, N., Rehm, E.,
Kallin, E., Kaiser, J., González- Posada, A., 2012. Estimates of net community
production and export using high-resolution, Lagrangian measurements of O2,
NO3−, and POC through the evolution of a spring diatom bloom in the North
Atlantic. Deep-Sea Res. I: Oceanogr. Res. Pap. 64, 157–174.
Andersson, M., LaCasce, J.H., Orvik, K.A., Koszalka, I., Mauritzen, C., 2011. Variability
of the Norwegian Atlantic Current and associated eddy field from surface
drifters. J. Geophys. Res. 116, C08032, http://dx.doi.org/10.1029/2011JC007078.
Antonov, J.I., Locarnini, R.A., Boyer, T.P., Mishonov, A.V., Garcia, H.E., 2006. World
Ocean Atlas 2005, Volume 2. In: Levitus, salinity.S. (Ed.), NOAA Atlas NESDIS 62.
US Government Printing Office, Washington, D.C. 182 pp.
Briggs, N., Perry, M.J., Cetinic, I., Lee, C.M., D'Asaro, E., Gray, A., Rehm, E., 2011. Highresolution observations of aggregate flux during a sub-polar North Atlantic
spring bloom. Deep-Sea Res. I: Oceanogr. Res. Pap. 58 (10), 1031–1039.
Dickson, B., Østerhus, S., 2007. One hundred years in the Norwegian Sea. Norsk.
Geografisk Tidsskr. 61 (2), 56–75, http://dx.doi.org/10.1080/00291950701409256.
Dinsmore, R.P., 1996. Alpha, Bravo, Charlie… Ocean Weather Ships 1940–1980.
Oceanus 39, 02.
Egbert, G.D., Erofeeva, S.Y., 2002. Efficient inverse modeling of Barotropic ocean
tides. J. Atmos. Oceanic Technol. 19, 183–204.
Eriksen, C.E., Osse, T.J., Light, R.D., Wen, T., Lehman, T.W., Sabin, P.L., Ballard, J.W.,
Chiodi, A.M., 2001. Seaglider: a long-range autonomous underwater vehicle for
oceanographic research. IEEE J. Oceanic Eng. 26 (4), 424–436.
Furevik, T., Nilsen, J.E.O., 2005. Large-scale atmospheric circulation variability and
its impacts on the Nordic Seas Ocean Climate—a review. In: The Nordic Seas:
An Integrated Perspective, Drange, H., Dokken, T., Furevik, T., Gerdes, R., Berger,
W. (Eds.), AGU Monograph 158, American Geophysics Union, Washington D.C.,
pp. 105–136.
Hansen, B., Østerhus, S., 2000. North Atlantic - Nordic Seas exchanges. Progress in
Oceanography 45 (2), 109–208.
F. Høydalsvik et al. / Deep-Sea Research I 79 (2013) 86–95
Hansen, B., Østerhus, S., Hátún, H., Kristiansen, R., Larsen, K.M.H., 2003. The
Iceland–Faroe inflow of Atlantic water to the Nordic Seas. Prog. Oceanogr. 59,
443–447.
Helland-Hansen, B., Nansen, F., 1909. The Norwegian Sea. Fiskeridir. Skr. Ser.
Havunders. 2, 1–360.
Isachsen, P.E., LaCasce, J.H., Mauritzen, C., Häkkinen, S., 2003. Wind-driven
variability of the large-scale recirculating flow in the Nordic Seas and Arctic
Ocean. J. Phys. Oceanogr. 33, 2534–2550.
Jakobsen, P.K., Ribersgaard, M.H., Quadfasel, D., Schmith, T., Hughes, C.W., 2003.
The near-surface circulation in the northern North Atlantic as inferred from
drifter data: variability from the meso-scale to interannual. J. Geophys. Res. 108,
3251, http://dx.doi.org/10.1029/2002JC001554.
Joyce, T.M., 1989. On in situ calibration of shipboard ADCPs. J. Atmos. Oceanic
Technol. 6, 169–172.
Locarnini, R.A., Mishonov, A.V., Antonov, J.I., Boyer, T.P., Garcia, H.E., 2006. World
Ocean Atlas 2005, Volume 1: temperature. In: Levitus, S. (Ed.), NOAA Atlas
NESDIS 61. US Government Printing Office, Washington, D.C. 182 pp.
Martin, J.P., Lee, C.M., Eriksen, C.C., Ladd, C., Kachel, N.B., 2009. Glider observations
of kinematics in a Gulf of Alaska eddy. J. Geophys. Res. 114, C12021, http://dx.
doi.org/10.1029/2008JC005231.
Mauritzen, C., 1996. Production of dense overflow waters feeding the North Atlantic
across the Greenland–Scotland Ridge. Part 1: evidence for a revised circulation
scheme. Deep-Sea Res. I 43, 769–806.
Mauritzen, C., Hansen, E., Andersson, M., Berx, B., Beszczynska-Möller, A., Burud, I.,
Christensen, K.H., Debernard, J., de Steur, L., Dodd, P., Gerland, S., Godøy, Ø.,
Hansen, B., Hudson, S., Høydalsvik, F., Ingvaldsen, R., Isachsen, P.E., Kasajima, Y.,
Koszalka, I., Kovacs, K.M., Køltzow, M., LaCasce, J.H., Lee, C.M., Lavergne, T.,
Lydersen, C., Nicolaus, M., Nilsen, F., Nøst, O.A., Orvik, K.A., Reigstad, M.,
Schyberg, H., Seuthe, L., Skagseth, Ø., Skarðhamar, J., Skogseth, R., Sperrevik,
A., Svensen, C., Søiland, H., Teigen, S.H., Tverberg, V., Wexels Riser, C., 2011.
Closing the loop—approaches to monitoring the state of the Arctic Mediterranean during the International Polar Year 2007–2008. Prog. Oceanogr. 90 (1),
62–89.
Mork, M., 1981. Circulation phenomena and frontal dynamics of the Norwegian
Coastal Current, Circulation and Fronts in Continental Shelf Seas. Royal Society,
London (UK), pp. 635–647.
Mork, K.A., Blindheim, J., 2000. Variations in the Atlantic inflow to the Nordic Sea,
1955–1996. Deep-Sea Res. I 47, 1035–1057.
Mork, K.A., Skagseth, Ø., 2010. A quantitative description of the Norwegian Atlantic
Current by combining altimetry and hydrography. Ocean Sci. 6, 901–911.
Oliver, K.I.C., Heywood, K.J., 2003. Heat and freshwater fluxes through the Nordic
Seas. J. Phys. Oceanogr. 33, 1009–1026.
Olsen, S.M., Hansen, B., Quadfasel, D., Østerhus, S., 2008. Observed and modelled
stability of overflow across the Greenland–Scotland Ridge. Nature 455 (7212),
519–522.
Orvik, K.A., 2004. The deepening of the Atlantic water in the Lofoten Basin of the
Norwegian Sea, demonstrated by using an active reduced gravity model.
Geophys. Res. Lett. 31, L01306, http://dx.doi.org/10.1029/2003GL018687.
95
Orvik, K.A., Skagseth, Ø., Mork, M., 2001. Atlantic inflow to the Nordic Seas: current
structure and volume fluces from moored current meters, VM-ADCP and
SeaSoar-CTD observations, 1995–1999. Deep-Sea Res. I 48, 937–957.
Orvik, K.A., Niiler, P.P., 2002. Major pathways of the Atlantic water in the northern
North Atlantic and Nordic Seas toward Arctic. Geophys. Res. Lett. 29 (19), 1896,
http://dx.doi.org/10.1029/2002GL015002.
Orvik, K.A., Skagseth, Ø., 2003. The impact of the wind stress curl in the North
Atlantic on the Atlantic inflow to the Norwegian Sea toward the Arctic.
Geophys. Res. Lett. 30 (17), 1884, http://dx.doi.org/10.1029/2003GL017932.
Orvik, K.A., Skagseth, Ø., 2005. Heat flux variations in the eastern Norwegian
Atlantic Current toward the Arctic from moored instruments, 1995–2005.
Geophys. Res. Lett. 32, L14610, http://dx.doi.org/10.1029/2005GL023487.
Østerhus, S., Turrell, W.R., Jonsson, S., Hansen, B., 2005. Measured volume, heat, and
salt fluxes from the Atlantic to the Arctic Mediterranean. Geophys. Res. Lett. 32,
L07603, http://dx.doi.org/10.1029/2004GL022188.
Perry, M.J., Sackmann, B.S., Eriksen, C.C., Lee, C.M., 2008. Seaglider observations of
blooms and subsurface chlorophyll maxima off the Washington Coast. Limnol.
Oceanogr. 53 (5, part 2), 2169–2179.
Poulain, P.-M., Warn-Varnas, A., Niiler, P.P., 1996. Near-surface circulation of the
Nordic Seas as measured by Lagrangian drifters. J. Geophys. Res. 101 (C8),
18,237–18,258.
Richter, K., Furevik, T., Orvik, K.A., 2009. Effect of wintertime low-pressure systems
on the Atlantic inflow to the Nordic seas. J. Geophys. Res. 114, C09006, http:
//dx.doi.org/10.1029/2009JC005392.
Rossby, T., Prater, M.D., Søiland, H., 2009. Pathways of inflow and dispersion of
warm waters in the Nordic seas. J. Geophys. Res. 114, C04011, http://dx.doi.org/
10.1029/2008JC005073.
Rossby, T., Flagg, C.N., 2012. Direct measurement of volume flux in the Faroe–
Shetland Channel and over the Iceland–Faroe Ridge. Geophys. Res. Lett. 39,
L07602, http://dx.doi.org/10.1029/2012GL051269.
Rudnick, D.L., Davis, R.E., Eriksen, C.C., Fratantoni, D.M., Perry, M.J., 2004. Underwater gliders for ocean research. Mar. Technol. Soc. J. 38, 73–84.
Saetre, R., Ljoen, R., 1972. The Norwegian Coastal Current. In: Proceedings of the
First International Conference on Port and Ocean Engineering, vol. 1, pp.
514–535.
Skagseth, Ø., Orvik, K.A., Furevik, T., 2004. Coherent variability of the Norwegian
Atlantic slope current determined by using TOPEX/ERS altimeter data. Geophys.
Res. Lett. 31 (14), L14304, http://dx.doi.org/10.1029/2004GL020057.
Søiland, H., Prater, M.D., Rossby, T., 2008. Rigid topographic control of currents in
the Nordic Seas. Geophys. Res. Lett. 35, L18607, http://dx.doi.org/10.1029/
2008JC005094.
Todd, R.E., Rudnick, D.L., Mazloff, M.R., Davis, R.E., Cornuelle, B.D., 2011. Poleward
flows in the southern California Current System: glider observations and
numerical simulation. J. Geophys. Res. 116, C02026, http://dx.doi.org/10.1029/
2010JC006536.
Trenberth, K.E., Large, W.G., Olson, J.G., 1990. The mean annual cycle in global ocean
wind stress. J. Phys. Oceanogr. 20, 1742–1760.
Worthington, L.V., 1970. The Norwegian Sea as a mediterranean basin. Deep Sea
Research and Oceanographic Abstracts, Elsevier, 77–84.
Download