OSOcean ScienceOSOcean Sci.1812-0792Copernicus PublicationsGöttingen, Germany10.5194/os-14-923-2018Moored observations of mesoscale features in the Cape Basin: characteristics
and local impacts on water mass distributionsMoored observations of mesoscale features in the Cape BasinKersaléMarionmarion.kersale@noaa.govLamontTarronhttps://orcid.org/0000-0001-8464-891XSpeichSabrinahttps://orcid.org/0000-0002-5452-8287TerreThierryLaxenaireRemihttps://orcid.org/0000-0001-5157-1821RobertsMike J.https://orcid.org/0000-0003-3231-180Xvan den BergMarcel A.AnsorgeIsabelle J.Marine Research Institute, Department of Oceanography – University
of Cape Town, Rondebosch, South AfricaCooperative Institute for Marine and Atmospheric Studies, University of Miami, Miami, Florida, USANOAA/Atlantic Oceanographic and Meteorological Laboratory, Miami, Florida, USAOceans and Coastal Research, Department of Environmental Affairs, Cape Town, South AfricaLaboratoire de Météorologie Dynamique, UMR 8539 École Polytechnique, ENS, CNRS, Paris, FranceIFREMER, Univ. Brest, CNRS, IRD, Laboratoire d'Océanographie Physique et Spatiale (LOPS), IUEM,
Plouzané, FranceOcean Science & Marine Food Security, Nelson Mandela University, Port Elizabeth, South AfricaMarion Kersalé (marion.kersale@noaa.gov)3September201814592394511October20171November201713July201816July2018This work is licensed under the Creative Commons Attribution 4.0 International License. To view a copy of this licence, visit https://creativecommons.org/licenses/by/4.0/This article is available from https://os.copernicus.org/articles/14/923/2018/os-14-923-2018.htmlThe full text article is available as a PDF file from https://os.copernicus.org/articles/14/923/2018/os-14-923-2018.pdf
The eastern side of the South Atlantic Meridional overturning circulation
Basin-wide Array (SAMBA) along 34.5∘ S is used to assess the
nonlinear, mesoscale dynamics of the Cape Basin. This array presently
consists of current meter moorings and bottom mounted Current and Pressure recording Inverted Echo
Sounders (CPIES)
deployed across the continental slope. These data, available from September
2014 to December 2015, combined with satellite altimetry allow us to
investigate the characteristics and the impact of mesoscale dynamics on local
water mass distribution and cross-validate the different data sets. We
demonstrate that the moorings are affected by the complex dynamics of the
Cape Basin involving Agulhas rings, cyclonic eddies and anticyclonic eddies
from the Agulhas Bank and the South Benguela upwelling front and filaments.
Our analyses show that exchange of water masses happens through the advection
of water by mesoscale eddies but also via wide water mass intrusions
engendered by the existence of intense dipoles. These complex dynamics induce
strong intra-seasonal upper-ocean velocity variations and water mass
exchanges between the shelf and the open ocean but also across the
subantarctic and subtropical waters. This work presents the first independent
observations comparison between full-depth moorings and CPIES data sets
within the eastern South Atlantic region that gives some evidence of eastern
boundary buoyancy anomalies associated with migrating eddies. It also
highlights the need to continuously sample the full water depth as
inter-basin exchanges occur intermittently and affect the whole water column.
Introduction
Mesoscale nonlinear dynamics contribute to large-scale water
mass distribution and therefore to the large-scale Meridional Overturning
Circulation (MOC) (Gordon, 1985; Biastoch et al., 2008; van Sebille et
al., 2012), as they redistribute momentum, heat and mass between different
regions (Robinson, 1983). Nonlinear mesoscale eddies, defined as nonlinear
using the advective parameter criteria of Chelton et al. (2011), are found
throughout the global oceans but are very energetic when associated with
western boundary currents. The Agulhas Current, the intense western boundary
current of the southern Indian Ocean, follows the African shelf edge
southwestwards along the east coast. When it reaches the southern tip of
Africa, at 34∘ S, the Agulhas Current leaves the continental slope
and affected by strong instabilities retroflects at 38∘ S into the
Indian Ocean (Lutjeharms and Cooper, 1996; de Ruijter et al., 1999;
Lutjeharms, 2006). These instabilities are responsible for extensive
meandering just southwest of South Africa, and shear edge features such as
Agulhas rings, eddies and filaments form within this region. These nonlinear,
mesoscale features control the Agulhas leakage – defined as the transport of
warm and salty Indian Ocean waters into the Atlantic Ocean through the Cape
Basin (labeled in Fig. 1; e.g., Gordon et al., 1992; de Ruijter et
al., 1999). The Agulhas leakage injects buoyancy anomalies that impact the
Atlantic MOC (AMOC) strength and the Atlantic Multi-decadal Oscillation, with
clear implications for climate (e.g., Beal et al., 2011; Biastoch et
al., 2015).
Agulhas ring observations, collected from shipboard surveys, were first
exploited to quantify the energy, heat and salt fluxes of these mesoscale
features (e.g., de Ruijter et al., 1999, and references therein). The first
inter-ocean exchange estimates were based on individual hydrographic cruise
data combined with altimetric tracking of Agulhas rings (Olson and Evans,
1986; Gordon and Haxby, 1990; Duncombe Rae, 1991; van Ballegooyen et
al., 1994; Byrne et al., 1995). These combined data revealed the presence of
four to eight rings per year propagating into the South Atlantic with speeds ranging
between 5 and 10 km day-1, with a diameter up to 400 km, and their
influence is felt from the surface down to an intermediate depth between 600
and 1100 m. The associated inter-ocean volume transport per ring was
estimated to be between 0.5 and 3 Sv (1Sv=106 m3 s-1).
These first estimates were based on statistics from the early constellation
of altimetry satellites that resolved fewer features than the modern
constellation and individual ring observations, which led to a large spread
of documented volume flux estimates.
A more accurate estimate of the fluxes in the Cape Basin has been
accomplished by deploying a line of instruments across the pathway of Agulhas
rings (BEST – Benguela Sources and Transport Experiment: Duncombe Rae et
al., 1996; Garzoli and Gordon, 1996; Goni et al., 1997) and deep profiling
floats (KAPEX – Cape of Good Hope Experiments: Lutjeharms et al., 1997;
Richardson et al., 2003; MARE – Mixing of Agulhas Rings Experiment: van Aken
et al., 2003). From these experiments, the northwestward propagation of the
rings in the Cape Basin was characterized through a narrow ring corridor
(Garzoli and Gordon, 1996) or through three different routes on the basis of
topographic effects (Dencausse et al., 2010). Such measurements have also
demonstrated the highly turbulent regime of the Cape Basin and the
coexistence of cyclonic and anticyclonic eddies which can enhance horizontal
mixing at intermediate depth (Boebel et al., 2003; Matano and Beier, 2003;
Richardson and Garzoli, 2003; Hall and Lutjeharms, 2011; Arhan et al., 2011).
Cyclonic eddies are about 50 km in diameter and are formed along the western
coast of South Africa or inshore of the Agulhas Current (Lutjeharms et
al., 2003). Tracked from altimetric data, recirculating plumes of warm water
at the sea surface are often identified around these eddies.
The presence of several Agulhas rings and cyclonic eddies can induce the
generation of dipole structures comparable to the Heton model of Hogg and
Stommel (1985). These oceanic dipoles have been observed using satellite
imagery thanks to their surface signature, described as a mushroom-like pattern
in various regions, and have been characterized with numerical simulations
and analytical theories (Ahlnäs et al., 1987; Hooker and Brown, 1994;
Hooker et al., 1995; Millot and Taupier-Letage, 2005; Pallàs-Sanz and
Viúdez, 2007; Baker-Yeboah et al., 2010a). These counterrotating eddies
enhance horizontal transport of heat along isopycnals, particularly across
frontal structures (Spall, 1995) and vertical diapycnal mixing. Related to
dipole dynamics, filaments extending well below the thermocline (de Steur et
al., 2004) have been observed during hydrographic cruises southwest of
Madagascar and in the Cape Basin (de Ruijter et al., 2004; Whittle et al., 2008).
Study area with shaded color representing the bathymetry (m) from
ETOPO1. The thin black line denotes the position of the SAMBA-east transect
and the black crosses (red squares) represent the mooring (CPIES) positions
with their associated numbers (1, 2, 3 and 4). The dotted black line denotes
the C-line position. Annually averaged 2015 SST isotherms from the ODYSSEA data set are plotted with
colored contours.
The mesoscale features in the Cape Basin can also interact with the eastern
boundary flow regime (Dencausse et al., 2010). The connection between
Agulhas rings and the South Benguela upwelling frontal zone has been
evidenced by the propagation of cold filaments from the coastal upwelling
system extending hundreds of kilometers offshore the Cape Peninsula
(Duncombe Rae et al., 1992; Lutjeharms and Cooper, 1996). Eventually joining
this coastal upwelling front, Agulhas filaments, originating from the
Agulhas Current, have been observed in the upper part of the water column
(Lutjeharms and Cooper, 1996). They have been associated with a strong
equatorward front jet, the Cape Peninsula jet (Bang, 1973; Gordon, 1986;
Lutjeharms and Meeuwis, 1987; Nelson et al., 1998). This strong equatorward
jet is located over the shelf-break, with a width of 20–30 km and can reach
a maximum velocity of 1.2 m s-1 (Shannon and Nelson, 1996; Veitch et
al., 2017). Its dynamics are suggested to be linked with the local wind
forcing during the upwelling season. However, this seasonality is still a
subject of discussion as Boyd et al. (1992) show the occurrence of the jet
beyond the upwelling season. This jet has been also linked to anticyclonic
circulations associated with frequent passages of Agulhas rings along the
continental slope (Dencausse et al., 2010). The issue is further complicated
by the regular generation of cyclonic eddies against the continental slope
in the area (Penven et al., 2001). These cyclonic eddies have been recently
associated with the intensification of a poleward undercurrent that is
generally weak to nonexistent in the Cape Basin (Baker-Yeboah et al., 2010b).
Moorings and CPIES characteristics. The date format is dd/mm/yyyy. The
bold values represent the depth of the instruments with pressure
recorders.
Understanding the Cape Basin circulation in the context of this strong
nonlinear, mesoscale variability is crucial for many studies. On regional
scales, the circulation is important for water mass distribution, local
dynamics, air–sea interactions and ecosystem assessments. For instance, the
interaction of Agulhas rings and filaments can have crucial implications for
cross-shelf exchanges and therefore productivity as water exported offshore
is generally rich in nutrients and biota, due to the upwelling regime and
the shelf edge jet transporting eggs and larvae alongshore. On large scales,
the Cape Basin circulation influences the AMOC and climate. The need of
observations in the subtropical South Atlantic at these different scales led to the establishment of the South Atlantic MOC (SAMOC) international
initiative to enhance further observing systems in this area (Garzoli and
Matano, 2011). As part of the SAMOC initiative, several efforts to document
the inflows of the two main paths of the upper limb of the AMOC have been
undertaken across the Drake Passage (e.g., Chereskin et al., 2009; Chidichimo et
al., 2014; Donohue et al., 2016) and south of South Africa (e.g., GoodHope line:
Ansorge et al., 2005; Speich et al. 2007; Gladyshev et al., 2008; Swart et
al., 2008; Hutchinson et al., 2016). A trans-basin AMOC array, South
Atlantic MOC Basin-wide Array (SAMBA), which began as a pilot array in
2008–2009 (Speich et al., 2010; Meinen et al., 2013), continues to grow
along 34.5∘ S, a crucial latitude to evaluate the MOC variability
and the impact of inter-ocean exchanges (Perez et al., 2011; Schiermeier,
2013). SAMBA is a collaboration between Argentina, Brazil, France, South
Africa and the United States, with moorings located on the western and
eastern boundaries (Speich et al., 2010; Meinen et al., 2013, 2018; Ansorge et
al., 2014). Since the pilot deployment described in
Meinen et al. (2013), the number of moorings on the boundaries has increased
dramatically. The transport and water mass anomalies associated with the
deep current and migrating eddies near the western boundary have been well
studied (e.g., Meinen et al., 2012, 2017; Valla et al., 2018), but
the eastern boundary anomalies have not yet been examined along
34.5∘ S.
In the framework of the SAMOC international project and national programmes
(SANAP – South African National Antarctic Programme), our work focuses on
the eastern part of the SAMBA array (hereafter SAMBA-east) offering an ideal
position to observe and characterize mesoscale dynamics – a key link between
Indian and Atlantic water exchanges (Fig. 1). The analysis of the SAMBA-east
moored data sets, from September 2014 to December 2015, provides evidence of
mesoscale features passing through the Cape Basin. Experimental evidence of
advected mesoscale eddies through moored arrays along a constant line of
latitude or longitude in various parts of the ocean have been presented
(Ursella et al., 2011; Sutherland et al., 2011; de Jong et al., 2014) and
used to estimate eddy parameters such as their strengths and sizes (Lilly and
Rhines, 2002). The main focus of this study is to characterize such mesoscale
structures in the Cape Basin region from different data sets obtained along
the SAMBA-east array using the Lilly and Rhines (2002) technique and to
quantify their impact on local water mass distributions.
Data and methods
The eastern part of the SAMBA array during September 2014 to December 2015,
consisted of four full-depth current meter moorings (hereafter full-depth
moorings), eight Current and Pressure recording Inverted Echo Sounder
(CPIES), and two bottom mounted acoustic Doppler current
profiler (ADCP) moorings that were deployed from the shelf to near the
Walvis Ridge offshore along 34.5∘ S (Fig. 1). In this study, we will
focus (1) on the four full-depth moorings (hereafter named M1, M2, M3 and M4)
extending from the continental shelf edge, at 1121 m of depth, to
15∘ E, at 4474 m of depth, deployed in September 2014 by the
Department of Environmental Affairs (DEA) and the University of Cape Town
(UCT), South Africa, from the R/V Algoa and (2) on the four CPIES
nearest to the full-depth moorings (hereafter named P1, P2, P3 and P4),
deployed in September 2013 by IFREMER, France, and DEA, South Africa, from
the R/V SA Agulhas II (Fig. 1, Table 1).
The characteristics of the moorings and CPIES used in this study are
summarized in Table 1. All full-depth moorings (M1–M4) have an
upward-looking 75 kHz RD Instruments ADCP, deployed at the uppermost float
at about 400–500 m of depth, set to sample the upper water column at hourly
intervals. At selected depths along the mooring lines, SBE 37 MicroCATs,
high-accuracy conductivity and temperature recorders, were deployed. Some of
these instruments also have pressure recorders and optical oxygen sensors.
The oxygen sensors will not be used in this study. All of the full-depth
mooring instruments (ADCP and SBE MicroCATs) were still recording on
recovery, with the exception of two SBE37 MicroCATs that stopped recording in
November 2014 due to low battery power. These two units were both from M4, at
depths of 1340 and 3985 m. The sampling period for the full-depth mooring
instruments was 1 h. The collected data were tidally filtered using a
second-order Butterworth filter with a 3-day cutoff period. CPIES recorded
hourly measurements of round-trip acoustic travel time, bottom pressure, and
the velocity at 50 m above the seafloor. All of the CPIES were recovered and
redeployed, with the exception of P3 which was unfortunately lost during the
recovery. Typically, the pressure sensor of the CPIES exhibits either a
linear or an exponential-plus-linear type of drift (Watts and Kontoyiannis,
1990; Donohue et al., 2010; Meinen et al., 2013). After removing the drift
following these traditional methods, the same tidal filter has been applied
on the CPIES data. The CPIES moorings records were subsampled to one value
per day at midnight UTC. An empirical look-up table for hydrographic property
profiles, known as the “gravest empirical mode” (GEM) method (Meinen
and Watts, 2000; Watts et al., 2001) was constructed from 2213
conductivity–temperature–depth (CTD) and Argo profiles in the region
(Fig. 2a). These two-dimensional look-up tables of temperature and salinity
as functions of depth and travel time are shown in Fig. 2b, c. The scatter
between the original hydrographic data and the GEM fields (Fig. 2d, e)
provides an estimate of the accuracy. Following the new thermodynamic
equation of seawater, all temperature and salinity is referenced as
conservative temperature (∘C) and absolute salinity (SA
(g kg-1)) (IOC et al., 2010). The temperature and salinity scatter is
less than 1.5 ∘C and 0.1 g kg-1, respectively. In the deep
ocean, the scatter around the GEM field is a larger fraction of the observed
variability; however, this is because the signals at depth are smaller
relative to the noise and not because the scatter around the GEM field itself
is significantly larger. The GEM method has been successfully applied in the
eastern South Atlantic (Meinen et al., 2013, 2018) and across the Agulhas
Current off the southeast coast of South Africa (Beal et al., 2015; Elipot
and Beal, 2018). Combining the measured travel time from the CPIES (Table 1)
with the GEM look-up tables produces daily full water-column hydrographic
profiles. We will primarily look at results from the full-depth moorings
(Sect. 3.1–3.3) and secondarily look at the CPIES moorings (Sect. 3.4). The
set of records presented hereafter were all tidally filtered as mentioned
above.
(a) Positions of the CTD and Argo casts from 1983 to 2018
collected around the SAMBA-east transect and used to create the GEM field.
The CPIES are identified by the red squares along the transect. Gravest
empirical mode (GEM) fields of conservative temperature (b) and
absolute salinity (c) determined for the SAMBA-east line. The
root-mean-squared (rms) differences between the original hydrographic
measurements and the smoothed look-up table values are shown in the lower
plots (d, e). The solid, dashed and dotted contours represent
progressively smaller contour intervals. The gray vertical lines show the
locations of the hydrographic measurements.
Following Lilly and Rhines (2002), a set of metrics used to characterize
eddies in mooring data were defined. The strength and size of an eddy can be
defined by the peak azimuthal velocity (V), the “apparent” radius or
half-width of the eddy chord passing by a mooring (X), the temporal center
of the eddy (t0), and the eddy's time influence on the mooring (ΔT). For an advected eddy, the
quantification of V needs a separation of the eddy flow from the mean
advection flow (magnitude U and direction Θ). The first estimate of
V (hereafter Vn) is made maximizing the velocity difference
over a time window encompassing the eddy and giving initial guesses for
t0 and ΔT. The mean advection flow direction Θ is then
estimated as it is expected to be perpendicular to the direction of
Vn. A second measure of V (hereafter VU) can be
obtained by subtracting an estimate of the advecting flow. For this, two
methods exist: the “angle” (Ua and Θa) and the
“filtering” (Uf and Θf) techniques. The first
method minimizes the angle difference between the observed eddy currents and
the observed vertical shear. The second method is useful when the advection
flow is large compared to the eddy and estimates Uf and
Θf by low-pass-filtered velocity time series at the eddy
center t0 using a Hanning window whose length is 3 times the eddy
duration ΔT. More details about these methods can be found in Lilly
and Rhines (2002) and the Appendix C of Lilly et al. (2003). For the purpose
of this study, the duration of the eddy (ΔT) and the magnitude of
the advection flow (Ua and Uf) are used to estimate
the apparent radius of the eddy: X=ΔT×U. Finally, the Rossby
number can be assessed as (2×|V|)/(X×f). In a purely linear
flow, the Rossby number is equal to zero and implies that the momentum
equation is dominated by the geostrophic balance between the pressure
gradient force and the Coriolis acceleration.
Hydrographic sampling has been conducted along the eastern section of the
SAMBA transect during five cruises on board the R/V SA Agulhas II
(September 2013, September 2014 and July 2015) and the R/V Algoa
(September 2014, and December 2015). During each cruise, CTD casts were
carried out at several stations along the transect using multiple SeaBird
Electronics SBE 911+ CTD systems and sensors, in order to measure
temperature and conductivity. Conductivity was used to derive salinity and
then, combined with temperature, to determine density. CTDs were conducted
from the surface to within 5–10 m of the seafloor on the R/V SA Agulhas II, with generally five casts offshore (0–15∘ E) and five
casts over the continental slope (15–17∘35′ E). CTDs on the R/V
Algoa were conducted to a maximum depth of 1000 m, with eight casts
over the continental slope (15–17∘35′ E) and a zonal resolution
of 0.3–0.5∘. Discrete seawater samples were collected at selected
depths for the analysis of salinity with a Guildline Portsal salinometer (van
den Berg, 2015) and used after the cruise to calibrate the CTD conductivity
sensors. Following the procedure of Kanzow et al. (2006), the SBE
MicroCat's sensors were attached to the rosette and data were recorded simultaneously
between the surface and 1000 m depth due to the limitation of the vessel. As
the CTD sensors on the R/V Algoa had not been recently calibrated by
the manufacturer and the surface mixed layer did not provide a stable
environment to assess the differences, the correction coefficients applied to
the MicroCat's data can introduce additional bias. To assess the accuracy of
these measurements, we compared the MicroCat's data with World Ocean Atlas
(WOA) 2013 climatology profiles (Locarnini et al., 2013; Zweng et al., 2013).
We compared the mean value of temperature and absolute salinity in the
MicroCat's records and the climatology profiles below 2000 m to avoid
depth-dependent offsets to the comparison due to strong vertical gradients in
the thermocline. The mean differences of 0.14 ∘C for the temperature
and -0.01 g kg-1 for the absolute salinity were estimated between
WOA and the MicroCATs below 2000 m. There differences are large – 1 or 2
orders of magnitude larger than the instrument accuracies – but they provide
an upper bound for the biases in the MicroCat's data in lieu of calibrated
CTD sensors. In any case, these mean differences are small compared with the
robust signals analyzed in this study.
Satellite imagery was analyzed to give an overview of the surface dynamics in
the Cape Basin region. Sea surface temperature (SST) was derived from
ODYSSEA, a Group for High Resolution Sea Surface Temperature (GHRSST)
regional product interpolated on a 0.02∘ grid for the South African
area. The sea surface height (SSH) and geostrophic velocity fields are
derived from the Delayed Time Maps of Absolute Dynamic Topography (MADT)
mapped daily on a 1/4∘ Mercator grid (Pujol et al., 2016). An eddy
detection method based on the algorithm developed by Chaigneau et al. (2008,
2009) was applied to these absolute dynamic topography (ADT) fields. This new
method detects local ADT extrema to identify potential eddy centers. For each
potential eddy centers, the algorithm looks for closed ADT contour lines with
an increment of plus or minus 1 mm (e.g., Chaigneau et al., 2011). This
1 mm threshold step is a bit smaller than the typical value used in the
literature; however, recent work has shown that it leads to more accurate
eddy sizes and amplitudes (Faghmous et al., 2015). This criterion prevents
the misidentification of large eddies as two smaller ones (Chaigneau et
al., 2008, 2009). The spatially largest closed ADT contour line encompassing
one extremum corresponds to the eddy edge. Following the filtering criteria
of Faghmous et al. (2015) for regional study, we only consider eddies with a
minimum lifetime of 7 days. The method provides the radius (Rout)
and the amplitude of the eddies associated with the outermost contour of ADT.
In addition, the eddy ADT contour along which the norm of the azimuthal
geostrophic velocity is the highest (Vmax) is calculated to
provide the radius of the coherent core of the eddy (RV; e.g., Nencioli
et al., 2010; Chelton et al., 2011). We validated our results with the
well-known Mesoscale Eddy Trajectory Atlas (META)
product based on an eddy-tracking methodology
developed by Chelton et al. (2011). The main difference between the META eddy
tracking and our methodology is the use of an improved eddy-tracking method
which takes into account both merging and splitting events. In both products,
the method is based on the low displacement of eddies compared to their size.
Each eddy was tracked within the boundary of the eddy defined at the previous
time step following the method of Pegliasco et al. (2015). In our version, no
cost function was applied to the eddies in cases of multiple association, but
a minimum of overlapping surfaces is added. Consequently, splitting events
are identified when one eddy is associated with two or more eddies the next
day as well as merging events in the reverse case. The eddies that are
discussed in the present paper are all identifiable by both the new method
and by the META product. In this study, the eddy statistics are based on our
method because it better tracks splitting/merging of eddies, and we only show
the META results to validate the new methodology.
Temporal evolution of the near-surface velocities recorded at the
last bin by the upward-looking ADCP (thin lines) and the surface geostrophic
velocity derived from satellite altimetry (bold lines). The
u (left column) and v (right column) component of the
current velocities (cm s-1) are represented at M1 (a, b –
blue line), M2 (c, d – green line), M3 (e, f – yellow
line) and M4 (g, h – red line). The correlation coefficient value
(R) is indicated in the top left corner of each panel.
ResultsUpper-ocean properties and dynamics
To characterize the hydrodynamical properties of the mesoscale features
passing through the SAMBA-east array, the measurements of the upward-looking
ADCP are analyzed together with concurrent altimetry data. From September
2014 to December 2015, the mean and the standard deviation of the near
surface velocities recorded at the shallowest depth (between 40 and 60 m
depth) by the upward-looking ADCP is between 24.6±11.4 cm s-1 at
the mooring nearest to the shelf (M1 – thin blue line; Fig. 3a, b) and
29.9±22.0 cm s-1 at the offshore mooring (M4 – thin red line;
Fig. 3g, h). Typically, the shallowest depth sampled was near the surface but
below the Ekman layer (mean value of about 36 m along the section) allowing
a comparison with the surface geostrophic velocity derived from satellite
altimetry. Statistical comparisons between ADCP near-surface velocities and the surface geostrophic
velocity derived from satellite altimetry have been made (Table 2, Fig. 3).
The comparison has been undertaken mostly in terms of correlation and rms
differences between the zonal and meridional components of M1–M4 ADCPs
(um, vm) and those from altimetry at the nearest gridded location
(ualti, valti):
ΔVrms=1N∑Vm-Vm‾-Valti-Valti‾21/2,
with V the zonal or the meridional components of the velocity and m the
index of the mooring. Additionally, comparisons were made by computing the
bias (Eq. 2).
bias=Vm‾-Valti‾
Correlations between ADCP and altimetry velocity estimates (R) are
significant and fell in the range 0.29–0.83 (Table 2). To determine the
significance of the correlations, the number of independent samples, also
known as the number of degrees of freedom (Thomson and Emery, 2014), is
estimated by dividing the record length by twice the integral timescale
calculated from the variability in each respective quantity (∼40–50 days). The integral timescale is estimated from the integral of the
autocorrelation function to its first zero crossing (see Appendix B in Meinen
et al., 2009 for more information). ΔVrms varied from 12.9
to 15.9 cm s-1 (Table 2). The absolute values of the biases were
typically between 1 and 6 cm s-1, with a maximum of 8.8 cm s-1
at M2. Correlation between the u-component of these velocity measurements
showed values greater than 0.7 for the three offshore moorings. Weak
correlation coefficients for both components were observed for the mooring
closest to the shore (M1, 1121 m of depth). The coefficient for the
v component gradually increases across the continental slope, moving away
from the shelf. The correlations for sites away from the shelf are
significant considering the mean zonal correlation scales of the satellite
product of 150 km at that latitude (Pujol et al., 2016). The poor
correlation for both velocity components at M1 can be mainly attributed to
its position too close to the coast (∼160 km off shore, at the shelf
break), and it is therefore embedded in a different dynamical regime than
purely geostrophic. The comparisons reveal that satellite data provide a
reasonable description of the upper-ocean circulation (between 40 and 60 m
depth) across the continental slope along our mooring arrays, except near the
shelf break (i.e., well inshore of the 1200 m isobath).
Summary of comparison statistics for satellite altimetry and moored
current meter, with R (the correlation coefficient),
ΔVrms (the root mean square differences between the zonal
and meridional components of ADCPs and those from altimetry (Eq. 1)) and the
bias between those components (Eq. 2).
RΔVrmsBias (cm s-1) (cm s-1) uvuvuvM10.310.2915.413.35.3-1.5(54.9 m)M20.720.5914.012.93.0-8.8(60.7 m)M30.810.6713.313.9-5.81.2(43.0 m)M40.800.8313.015.9-4.3-4.9(45.8 m)
Eddy trajectories from 15 September 2014 to 15 December 2015 for
eddies influencing the mooring measurements. The eddies are named A1–A7
(anticyclonic eddies – a) and C1–C6 (cyclonic eddies –
b), with the numbers assigned in chronological order by the
eddy-tracking scheme. The circles indicate the starting positions of an eddy track. The
bold blue line in (b) represents the trajectory
of an eddy after a merging. (a) The dashed black line represents the
northern route of Agulhas rings defined by Dencause et al. (2010).
Characteristics of the eddies passing over the moorings from
satellite altimetry. The time period, the radius, and the azimuthal velocity
of the different eddies are derived from our eddy detection method. The
eddies are named C1–C6 (cyclonic eddies) and A1–A7 (anticyclonic eddies),
with the numbers assigned in chronological order by the
eddy-tracking scheme.
(a) Vertical scheme of the mooring lines positions along
the latitude 34.5∘ S. The squares represent the upward-looking ADCP
positions, the small horizontal lines indicate the SBE MicroCAT's sensors and
the circles at the bottom show the CPIES. Temporal evolution of the current
vector stick plot at M2 (b), M3 (c) and M4 (d).
One vector per day is shown. The shaded areas show the eddy events defined
with altimetry data (red: anticyclonic eddies; yellow: dipole event; blue:
cyclonic eddies).
In light of these results, the dynamics inferred from satellite altimetry can
provide a basis for understanding the variability in the zonal and meridional
components of the upper-layer velocity between 40 and 60 m depth for the
offshore moorings (M2 to M4). Using the new eddy detection technique, we can
first estimate statistical characteristics of mesoscale eddies passing
through the mooring line during the measurement period of the full-depth
moorings along 34.5∘ S (September 2014 to December 2015). A total
number of 16 Agulhas rings, defined as anticyclones that enter the Cape Basin
crossing the C-line (Fig. 1), were detected and confirmed by the META
product. This line extends from the southernmost tip of Africa (Cape Agulhas)
and, after crossing various seamounts, ends at 45∘ S in the Southern
Ocean. From our altimetric tracking, the median radius and standard deviation
of these features are equal to 85±43 km for Rout and 66±38 km for RV. The azimuthal speed of these Agulhas rings
(Vmax) is equal to 0.49±0.24 m s-1 with a translation
speed, mainly northwestward of 11±6 km day-1. Considering our
observing system along the SAMBA-east line, seven anticyclonic and six
cyclonic eddies influenced the mooring measurements during the ∼14-month period of recording at sites M2 (P2) to M4 (P4) (Fig. 4, Table 3).
All these mesoscale features passed over one of the moorings considering
their outermost contours of ADT and have a closed contour in the satellite
dynamic-height amplitude of at least 1 cm. Two of these eddies (anticyclonic
A4 and cyclonic C4) were generated at the Agulhas retroflection and
propagated into the Cape Basin through the northern route defined by
Dencausse et al. (2010) (dashed black line – Fig. 4a). These two eddies have
the largest azimuthal speed (Vmax) compared with other eddies
detected in the area (Table 3). Four of the anticyclonic eddies (A1, A2, A5
and A7) observed in this study are generated by the splitting of an Agulhas
ring and two are generated north of the SAMBA line (A3 and A6). Cyclonic eddy
C5 is generated by the splitting of C4 and the other four cyclonic eddies are
generated over the slope off the Cape Peninsula.
The currents measured during September 2014–December 2015 in the top 500 m
of the water column contain a number of sudden rotational events, seen in the
time series of the current vector stick plot (Fig. 5). These transitions in
the velocity field are associated with mesoscale eddies passing through the
mooring line. The altimetry data allow us to examine the impact of six
cyclonic (blue shaded areas – Fig. 5) and seven anticyclonic eddies (red
shaded areas – Fig. 5) on zonal and meridional velocity at the three mooring
sites. Moreover, the data records show that the presence of one dipole
(counterrotating eddies) affects the velocity measurements (yellow shaded
area – Fig. 5). The presence of this dipole is confirmed in the ODYSSEA SST
field by its mushroom-like pattern, the typical surface signature of these
features.
To have more of a consistent and complete picture of the mesoscale dynamics
in the Cape Basin, the position of these eddies observed by satellite
altimetry were analyzed in relation to variations in the ODYSSEA SST field
(Fig. 6). On 1 December 2014, both eddy identification methods detect one
cyclonic eddy (C1) and one anticyclonic eddy (A1) close enough to the mooring
line to affect the measurements (Fig. 6a). Approximately 1 month later
(Fig. 6b), C1 and A1 move westward. At that time, a new cyclonic eddy (C2) is generated on the slope, at the
South Benguela upwelling front, and a new anticyclonic eddy (A2) is generated by the splitting of
A1.
On 20 February 2015 (Fig. 6c), an anticyclonic eddy (A3) is present between
M3 and M4 and a new cyclonic eddy (C3) is generated at 35∘ S. This
cyclonic eddy merges afterwards with an intense cyclonic eddy (C4) coming
from the Agulhas retroflection. After the merging, the cyclonic eddies C3 and
C4 (called C4)
affect the measurements of M2–M3 until the end of March (Fig. 5c, d). This
strong northeastward current is also intensified by an intense cross-shelf
density front that is enhanced at that time due to an upwelling event and the
northward migration of an Agulhas ring A4 (Fig. 6d). On 21 March 2015, A4
splits to generate A5. This anticyclonic eddy was close enough to M4 to
induce the northward propagation of a warm filament on 9 April 2015
(Fig. 6e).
SST satellite images for selected dates between 1 December 2014 to
15 September 2015. The black line denotes the position of the SAMBA line, and
the crosses represent the mooring positions. Black contours show the coherent
core of the eddies identified by the new method of detection. The colored
dots (red for anticyclonic eddy and blue for cyclonic eddy) are the eddy centers detected in the META product.
At the end of April, an intense dipole is observed due to the interaction of
eddies A5 and C5. Note that cyclonic eddy C5 is generated by the splitting of
C4 on 18 April 2015. The dipole (A5, C5) induces a strong northward current that injected cold surface water
across the mooring array on 25 April (Fig. 6f) and warmer Agulhas Current
water during the month of May (Fig. 6g).
After this intense event, the satellite imagery still reveals the presence of
cyclonic eddy C5 and a new anticylonic eddy (A6) in July between M3 and M4
(Fig. 6h). On 15 September 2015 (Fig. 6i), the presence of a cyclonic eddy
(C6) over the mooring line and an anticyclonic eddy (A7) generate a warm
filament propagating toward M4.
In summary, the moorings over the slope (M2 and M3) are affected by five
cyclonic eddies identified by the two methods of eddy detection (Table 3;
Figs. 5 and 6) over the measurement period from September 2014 to December
2015. The largest velocity perturbations over the slope are associated with
cyclonic eddy C4. M4 records show the influence of many anticyclonic eddies
and three cyclonic eddies. The largest velocity perturbations are seen at M4
during the presence of the dipole A5 and C5. The presence of several of these
mesoscale features induces the propagation of two warm filaments and two
injections of relatively colder surface water and relatively warmer Agulhas
Current water.
(a, b, c) Hodograph of the mean currents of the upper water
column observed at M4 for three case studies (a – anticyclonic
eddy; b – cyclonic eddy; c – dipole). Color represents
the temperature recorded at the first SBE37 MicroCATs
de-meaned over the time period
of each case studies (a: 9 March–6 April 2015; b:
28 May–10 June 2015; c: 8 April–29 May 2015). The estimated
centers of six events have been marked with white stars. The beginning of
each time series has been marked with white dots.
(d, e, f) Progressive vector diagram for the mean currents of the
upper water column observed at M4 for three case studies. Color, time period
and symbols are the same as the hodograph.
Case studies – upper water column
From the combined analysis of satellite and mooring data in Sect. 3.1, we
selected the period between March and June 2015 to analyze eddy-like
features, filaments and strong intrusions of cold water or Agulhas Current
water due to eddy–eddy interactions. We focus on the variability at M4,
which exhibits the strongest variations in responses to passing mesoscale
features.
Following the method of Lilly and Rhines (2002), the coherent eddies, dipoles
or filaments in mooring data can be detected and assessed in a more
quantitative fashion. The zonal and meridional components of the velocity
measured at the mooring can be placed on a hodograph plane (Fig. 7a, b, c).
In this plane, a combination of a straight line with a segment of a circle
(called D-shaped) reflects an eddy structure. The nearer the eddy's center
passes to the mooring, the more the hodograph appears as a straight line. An
eddy sliced through its exact center or a filament lead to the same type of
hodograph (straight line). Eddies can be further distinguished from fronts
and filaments using progressive vector diagrams (PVDs) (Fig. 7d, e, f). The
eddy structure presents bends, unlike fronts or filaments, which result in
straight lines. Concerning the hodographs for a dipole, it is similar to that
of an eddy if one of the cores passes over the mooring. If the mooring
measures the velocity in between the two cores, the hodograph shows a pulse
with a strong increase and decrease in the currents.
In accordance with the Lilly and Rhines (2002) detection method, we have
isolated three case studies between March and June 2015. Each case study is
associated with one feature which can generate different events. The first
case study focused on anticyclonic eddy A4, its splitting which generates A5
and the propagation of a filament along its border. The second case study
detailed the dynamics of cyclonic eddy C5, and the third one described the
dynamics of a dipole (A5/C5) generating two current pulses.
Estimate eddy parameters. t0: the eddy's center apparent at M4;
ΔT: temporal half eddy duration; U/Θa/f
(Xa/f): the estimated mean flow magnitude and direction (eddy
apparent radius) using the angle method (subscript “a”) or the filtering
method (subscript “f”). Vn/VU: the estimated maximum
eddy currents using the normal (subscript “n”) or the U subtraction method
(subscript “U”). Rossby number using the different value for V and X.
A4A5C5t014 Mar 2015 – 15:0024 Mar 2015 – 21:003 June 2015 – 16:00ΔT (h)544825Ua (cm s-1)27.018.350.7Θa110772Uf (cm s-1)28.916.549.5Θf150772Vn (cm s-1)25.820.6-19.2VU (cm s-1)26.021.7-20.0Xa (km)52.431.645.7Xf (km)56.228.444.5Rossby number0.11–0.120.15–0.170.10
Anticyclonic eddy: (a, b) SST satellite images for selected
dates (vertical black lines on the time series). Vertical–temporal section
of the u component (c) and v component (d) of the
current speed measured from the upward-looking ADCP at M4. The colored
circles at the top of the section show the u and v component of the
current speed from altimetry data. (e) Temporal evolution of
temperature (blue line) and salinity (green line) and (f) the
isopycnal displacement recorded at the first SBE37 MicroCATs between 430 and
450 m depth.
The first case study between 9 March and 6 April 2015 (Fig. 7a, d)
illustrates the presence of anticyclonic eddies A4 and A5. The D-shape
structures on the hodograph (Fig. 7a) and the bends to the right associated
with anticylonic eddies on the PVD (Fig. 7d) appear clearly. To better
illustrate these events, the temperature recorded at the first SBE MicroCATs
(depth between 430 and 750 m) de-meaned over the time period of each case study is
presented. Within the period during which a positive temperature anomaly
associated with the anticyclonic eddies appears in the record, the hodograph
tends to be relatively straight. At the end of the time period, the impact of
a filament propagating northward due to the presence of A5 is observed as a
straight line on both planes (Fig. 7a, d). The second hodograph and PVD
between 28 May and 10 June 2015 (Fig. 7b, e) display a D-shaped structure and
a bend to the left associated with a negative temperature anomaly typical of
a cyclonic eddy. According to its time of occurrence, this mesoscale feature
corresponds to C5. Finally, the final case study between 8 April and 29 May
2015 (Fig. 7c, f) exhibits very strong currents with two acceleration and
deceleration phases that are associated with a dipole.
Following these three identified cases studies, the vertical–temporal
structure of eddy-like features and filaments and strong intrusions of cold
water or Agulhas Current water due to eddy–eddy interactions are analyzed.
Moreover, the estimated sizes and strength of eddies (A4, A5 and C5) can be
defined (Table 4) according to the Lilly and Rhines (2002) method detailed in
Sect. 2.
Cyclonic eddy: (a) SST satellite
image for selected date (vertical
black line on the time series).
Vertical–temporal section of the u component (b) and
v component (c) of the current speed measured from the
upward-looking ADCP at M4. The colored circles at the top of the section show
the u and v component of the current speed from altimetry data.
(d) Temporal evolution of temperature (blue line) and salinity
(green line) and (e) the isopycnal displacement recorded at the
first SBE37 MicroCATs between 470 and 500 m depth.
Intrusion water: (a, b) SST satellite images for selected
dates (vertical black lines on the time series). Vertical–temporal section
of the u component (c) and v component (d) of the
current speed measured from the upward-looking ADCP at M4. The colored
circles at the top of the section show the u and v component of the
current speed from altimetry data. (e) Temporal evolution of
temperature (blue line) and salinity (green line) and (f) the
isopycnal displacement recorded at the first SBE37 MicroCATs between 435 and
700 m depth.
(a) Conservative temperature (∘C)–absolute
salinity (g kg-1) relationship from the hydrographic cruises sampling
along the SAMBA-east array on board the RV SA Agulhas II (black
dots) and from the SBE37 MicroCAT measurements at M4 (depth – colored
circles). The gray contours represent the potential density anomaly with a
reference pressure of 1000 dbar. Water masses are indicated by gray boxes
(OSW – Oceanic Surface Water; MUW – Modified Upwelled Water; lSACW – Light
South Atlantic Central Water; SASTMW – South Atlantic Subtropical Mode
Water; SAMW – Subantarctic Mode Water; ARMW – Agulhas Ring Mode Water;
I-AAIW – Indian Antarctic Intermediate Water; IA-AAIW – Indo-Atlantic
Antarctic Intermediate Water; A-AAIW – Atlantic Antarctic Intermediate
Water; UCDW – Upper Circumpolar Deep Water; NADW – North Atlantic Deep
Water; LCDW – Lower Circumpolar Deep Water). (b, c) Close-up views
of the deep water recorded with the MicroCat illustrating the distributions
observed in the anticyclonic/cyclonic cases (b) and intrusion
cases (c). The colors represent the measurements during each case
study(anticyclonic eddy: red; cyclonic eddy: blue; intrusions: cold water –
green, warm water – orange).
Vertical profiles of the conservative temperature (a, c)
and absolute salinity (b, d) anomalies at M4 due to the presence of
the anticyclonic and cyclonic eddies (a, b) and the two intrusions
associated with the dipole (c, d). The colored dots represent the
anomalies from the SBE37 MicroCATs with their estimated accuracy (horizontal
black line) and their range of neutral density variability (colored boxes).
The solid colored lines represent the anomalies from the reconstructed GEM
fields. The gray shaded areas around the horizontal zero axis show the range
of value below the estimated measurement's accuracy.
The case of anticyclonic eddies. From the middle to the end of March
(Fig. 8a, b), M4 is impacted by Agulhas rings A4 and A5. The
vertical–temporal section of the u and v component of the current speed
(Fig. 8c, d) shows an eddy-like structure on 14 and 24 March 2015. The
vertical structure of the first Agulhas ring A4 is not evident as it is
associated with an advecting flow directed to the southwest (Θ –
Table 4). For the second anticyclonic eddy, A5, the direction of the mean
flow shifts toward the west (Θ – Table 4) and so the eddy-like
structure is much clearer in the perpendicular direction (v component,
Fig. 8d). The azimuthal eddy velocities reach 26.0 cm s-1 for A4 and
21.7 cm s-1 for A5 (VU – Table 4) from the surface to
200–250 m depth. During this period, an increase in temperature (salinity)
of 2.4 ∘C (0.35 g kg-1) is recorded at the first SBE37
MicroCATs between 430 and 450 m depth (Fig. 8e). We can also evaluate a
downshift of the isopycnal layer by about 200 m for both eddies (Fig. 8f).
This quantity is found by interpolating the potential density within the mean
vertical profile derived from CTDs on the R/V SA Agulhas II. At the
end of the time period (2 April), the propagation of a warm filament at the
eastern side of A5 is identified. The measured v component of the velocity
(Fig. 8d) shows at that time a strong northward flow. From the satellite
altimetry (colored circles at the top of Fig. 8c, d), the velocity component
magnitudes are lower than the ones from the upward-looking ADCP but show
changes of a similar order.
The case of a cyclonic eddy. The same analysis is undertaken for a
second event occurring between 28 May and 10 June 2015. On 3 June 2015
(Fig. 9a), a cyclonic eddy (C5) affects the circulation along the SAMBA line.
At this date, the vertical–temporal section of the u component of the
current speed (Fig. 9b) shows an eddy-like structure. During the time period
the eddy passes across the mooring, a strong northward flow is recorded
(v component – Fig. 9c; Θ and U – Table 4). The core of this
strong cyclonic eddy (-20.0 cm s-1 azimuthal velocity) extends from
the surface down to 200 m depth. A decrease in salinity and temperature
between 470 and 500 m depth of 0.13 g kg-1 and 1.7 ∘C,
respectively, is recorded (Fig. 9d). This decrease can be partly explained by
the 100 m uplift of the isopycnal layer (Fig. 9e). From the satellite
altimetry (colored circles at the top of Fig. 9b, c), the meridional
component of the current speed is in good agreement with the upward-looking
ADCP compared to the zonal component that shows a lower
velocity.
Intrusions – dipole dynamics. From the middle of April to the end
of May (Fig. 10), anticyclonic eddy A5 and cyclonic eddy C5 affect the
circulation around M4 as a dipole. This intense dipole induces two intense
northward pulses of current with meridional velocities exceeding
100 cm s-1 on 22 April and 14 May throughout the entire depth of the
sampled water column. These two pulses last about 22 and 15 days. Their
vertical influence is deeper than 600 m depth. The vertical mooring motion
as evidenced by the downward shifts in the range of depths resolved by the
ADCP is very intense during these two events. Although depth changes during
these events (∼300 m) were substantial, the minimal variations in
pitch and roll (maximum change of 5 and 4∘), were well within
accepted limits and indicated that the performance of the mooring is
satisfactory. The altimetry data show the same intense northward current at
the surface; however, there is a pronounced lag of 7 days for the second
intrusion. This lag is no longer present if we consider the altimetric
velocity one grid eastward (0.25∘ of resolution). These two pulses of
currents impact the measurements at the first SBE MicroCAT's sensor between
435 and 700 m depth (Fig. 10e, f) with large variations in temperature (more
than 6 ∘C) and salinity (∼0.7 g kg-1) and a downshift
of isopycnal layers by about 150–200 m.
Note that the temperature and salinity anomalies for all three of these cases
are much larger than the upper bound of the MicroCat's temperature and
salinity biases (0.14 ∘C and 0.01 g kg-1), documented in
Sect. 2.
Table with the different water masses and their definition in terms
of conservative temperature, absolute salinity and density layers.
Water mass DefinitionMUW 9.71≤T≤21.3∘C; 34.661≤SA≤35.647 g kg-1OSW γn<26.2 kg m-3; T≥16.0∘C; SA≥35.647 g kg-1lSACW 15.15≤T≤16∘C; 35.506≤SA≤35.764 g kg-1SASTMW 11.69≤T≤16∘C; 35.49≤SA≤35.764 g kg-1SAMW 6.32≤T≤13.18∘C; 34.78≤SA≤35.49 g kg-1ARMW 11.60≤T≤17.00∘C; 35.18≤SA≤35.81 g kg-1I-AAIWSA≥34.47 g kg-1AAIWIA-AAIW34.37≤SA≤34.47 g kg-1A-AAIWSA≤34.37 g kg-1UCDW 27.55<γn<27.92 kg m-3; T<4.23∘C; 34.696≤SA≤35.916 g kg-1NADW 27.92<γn<28.11 kg m-3LCDW 28.11<γn<28.26 kg m-3Full water-column water mass distribution and variability
We used the daily averaged temperature and salinity data obtained from SBE37
MicroCat's instruments on moorings to recover the regional water masses
present at M4. Similarly to Lamont et al. (2015), the distribution of water
masses was determined according to conservative temperature, absolute
salinity and density layers, as illustrated in Fig. 11a and described in
Table 5. Modified Upwelled Water (MUW) was defined according to Duncombe Rae
(2005) as central shelf water upwelled along the coast and modified due to
solar heating and freshwater flux. Oceanic Surface Water (OSW) was defined
with the criteria of Donners et al. (2005) as salinity maximum water
subducting in the western tropical Atlantic. The criteria of Donners et
al. (2005) were also used to define light South Atlantic Central Water (lSACW
– defined as Indian Central Water brought into the South Atlantic Ocean by
Agulhas Current intrusions), South Atlantic Subtropical Mode Water (SASTMW)
and Subantarctic Mode Water (SAMW – with a vertical temperature gradient
less than 1.6 ∘C per 100 m; Roemmich and Cornuelle, 1992). Local ventilation of Indian
Central waters has been firstly identified by Arhan et al. (2011) inside
different sampled Agulhas rings (Arhan et al., 1999; Gladyshev et al., 2008).
This water mass has been recently defined in the study of Capuano et
al. (2018) as Agulhas Ring Mode Water (ARMW) including in addition to the
above two other Agulhas rings previously sampled (Duncombe Rae et al., 1996;
McDonagh et al., 1999). Three different varieties of Antarctic Intermediate
Water (AAIW), namely Indian AAIW (I-AAIW), Indo-Atlantic AAIW (IA-AAIW) and
Atlantic AAIW (A-AAIW), were characterized according to Rusciano et
al. (2012). The highest salinity values in the I-AAIW variety are likely
associated with Red Sea Intermediate Water (RSIW), which has been shown
traveling down the Agulhas current as discontinuous filaments or confined
within anticyclonic and cyclonic eddies (Roman and Lutjeharms, 2007). Upper
Circumpolar Deep Water (UCDW), North Atlantic Deep Water (NADW) and Lower
Circumpolar Deep Water (LCDW) were defined according to Heywood and King
(2002).
Overall, the vertical distribution of water masses at M4 (colored dots –
Fig. 11a) shows that SAMW is present on the southwestern African continental
slope around 500 m depth, I-AAIW and IA-AAIW between 500 and 1000 m, UCDW
from 1000 to 1500 m depth, NADW between 1600 and 3000 m, and finally LCDW
below 3000 m.
The mean vertical and zonal distributions are affected by the regional
mesoscale dynamic described in the previous section. Typically a cyclonic
(anticyclonic) eddy causes uplift (suppression) of isopycnal layers. The
temperature and salinity relationship is highlighted for these two types of
events. During the period of the anticyclonic eddies, the temperature and
salinity and density values at the shallowest SBE MicroCAT sensor at M4 (red
dots – Fig. 11b) show the highest range of values of the all of the time
series records. During the time period of the cyclonic eddy, the values are
at the opposite end, i.e., the lowest recorded densities (blue dots –
Fig. 11b). While the signature of these two features is clearly separated for
the two shallowest SBE MicroCAT sensors, we do not definitively prove whether
these changes are associated with a thermohaline anomaly or a simple heave.
During the two successive intrusions of waters due to the presence of the
dipole at M4, the temperature and salinity relationship (orange and green
dots – Fig. 10c) highlights a wide range of values. For this type of event
the deepest sensors recorded water masses characteristics different from the
ones usually sampled over the time period (Fig. 11a). The pressure record
from the sensor at M4 exhibits the largest vertical variation during the time
period of the dipole (Fig. 10c, d). This large vertical mooring motion adds
another level of complexity to interpreting this relationship. To understand
the origin of these variations, which can be associated with water trapped
inside or around the eddies, vertical movement of isopycnal layers, and/or
mooring motions, the full water column characteristics will be analyzed with
different data sets on neutral density surfaces.
Full water-column analysis – focus on the three case studies
The different data sets allow us to analyze in more detail the effects of
the mesoscale features described in the previous sections and successfully
cross-validate the measurements made by the different types of moored
instruments (full-depth moorings vs. CPIES).
The profiles estimated via the GEM method can be compared with the single
point values of temperature and salinity from M4 sensors (eight sensors
recording over the common time period, from 20 September 2014 to 11 August
2015). The reconstructed field captures the major changes in temperature and
salinity variability in the upper water column. Indeed, significant
correlation coefficients between these two independent data sets show values
higher than 0.93 for salinity and temperature at the shallowest sensors
(∼450, 900 m depth). At the deepest levels, the correlation
coefficients range between 0.14 and 0.83, and they are not significant at the
95 % confidence level. This can be due to the small variability at the
deepest levels compared to the scatter around the GEM field (Fig. 2) and the
correction applied to some of the MicroCAT's data without pressure recorder
during intense vertical mooring motion.
Measurements from these two data sets are compared for the three case studies
described in Sect. 3.2. The conservative temperature and absolute salinity
anomalies at M4 due to the presence of the eddies (anticyclone and cyclone:
A5 and C5; Fig. 12a, b) and the intrusion of water that arise because of the
presence of the dipole (Fig. 12c, d) are calculated relative to the
hydrographic properties in “normal conditions” (just before each event
occurred).
During the occupancy of A5 at M4, defined as the temporal range from t0-3ΔT to t0 (Table 4 – 18–24 March 2015), the measurements at the
SBE37 MicroCATs show an increase in temperature of 1.54 ∘C and
salinity of 0.22 g kg-1 along the 26.73 kg m-3 neutral surface
(Fig. 12a, b – red dots). The second sensor recorded a temperature anomaly
of 0.48 ∘C. No significant anomaly is recorded for the salinity
along the 27.33 kg m-3 neutral surface. The reconstructed temperature
profile via the GEM method in the upper part of the water column captures a
maximum temperature anomaly of 1.42 ∘C along the 27.03 kg m-3
neutral surface and a salinity of 0.17 g kg-1 along the 26.7 kg
m-3 neutral surface (Fig. 12a, b – red line), which is relatively well
captured with the single-point measurements of the full-depth mooring.
Negligible anomalies are detected in the deeper part of the water column with
a neutral density larger than 27.7 kg m-3.
During the time period of C5, defined as before from t0-3ΔT to
t0 (Table 4 – 31 May–3 June 2015), the shallowest SBE37 MicroCat records
negative temperature (-1.43∘C) and salinity
(-0.17 g kg-1) anomalies (Fig. 12a, b – blue dots). Along the
27.55 kg m-3 neutral surface, the second sensor records saltier water.
Anomalies of 0.04 g kg-1 for the salinity are characteristics of
I-AAIW. The hydrographic data, estimated from the CPIES, show anomalies
associated with this feature of the same order. The transition between the
different sign of salinity anomalies is evidenced at 27.35 kg m-3.
Negligible anomalies are detected in the deeper part of the water column.
For both features (cyclonic and anticyclonic eddies), the similar temperature
and salinity anomalies between the two data sets in the upper water column
(26–27.7 kg m-3) provide evidence for the presence of different water
masses trapped inside the observed mesoscale features.
During the two intrusions of water due to the presence of the dipole (first
intrusion: 11–22 April 2015; second intrusion: 7–14 May 2015 – green and
orange dots, Fig. 12c, d), the temperature and salinity show large anomalies
at all neutral densities sampled. The largest anomalies are recorded at the
first SBE37 MicroCATs for both intrusions. During the second intrusion, a
warmer (ΔT=4.03∘C) and saltier (ΔS=0.41 g kg-1) water of Indian origins appears in the upper part of the
water column (<27 kg m-3 neutral surface). Negative anomalies
during the first intrusion are very likely associated with water of
subantarctic origins (anomalies of -5.19∘C for temperature and
-0.56 g kg-1 for salinity). As for the cyclonic eddy, during the
cold intrusion the presence of high-salinity I-AAIW (0.28 g kg-1 in
salinity) is observed along the 27.52 kg m-3 neutral surface.
Interestingly from the SBE37 MicroCATs, we observe strong anomalies of
salinity at the lowest neutral surface (>28 kg m-3). The presence
of such saltier (ΔS=0.16 g kg-1) water anomalies at these
depths reveal the presence of large pulses of NADW. This event influences not
only the upper-layer waters but its also affects most of the water column,
down to the bottom.
From the reconstructed CPIES fields, the temperature and salinity signatures
of the two intrusions (Fig. 12c, d – green and orange lines) show anomalies
of the same sign but 2 times smaller than the ones recorded at the
MicroCATs. During these intrusions, the vertical mooring motion is very
intense and is larger than the isopycnal displacement. This result is
supported by the large range of neutral densities sampled by the SBE37
MicroCATs (colored boxes, Fig. 12c, d). This range is even larger for the
first two sensors, as at these shallow depths the vertical gradients of
salinity and temperature are larger than horizontal gradients.
Discussion and concluding remarks
Since 2010, several efforts
have been undertaken to further enhance the AMOC observation in the South
Atlantic. This strategic monitoring system continues to grow along
34.5∘ S, a crucial latitude to evaluate the AMOC variability and the
impact of inter-ocean exchanges (Drijfhout et al., 2011). As a consequence of
the limitations of high spatial and temporal resolution in in situ
observations, the quantification of inter-ocean exchange is still ongoing
work and many key questions and issues remain open, such as what the
characteristics of mesoscale structures are and what their impact on local
water mass exchange and distribution is. In the framework of the SAMOC
initiative, we provide here further investigation using a combination of
satellite altimetry, full-depth moorings measurements and CPIES records.
Focusing on the SAMBA-east array region, the general circulation around South
Africa has been rather well described in previous studies. Two main processes
have been observed to influence this area: an equatorward shelf break frontal
jet off the Cape Peninsula (Lutjeharms and Meeuwis, 1987; Nelson et
al., 1998) and the instabilities of the Agulhas Current responsible for the
spawning of mesoscale eddies propagating into the Cape Basin (Lutjeharms and
Cooper, 1996; de Ruijter et al., 1999; Lutjeharms, 2006).
During the time period of our study (∼14 months), we used a newly
developed method to identify a number of eddies in our region using satellite
imagery; these eddies were then confirmed to also be present in the
well-known META atlas (Chelton et al., 2011). Further analysis showed that 16
large eddies were Agulhas rings, and the satellite data show that one of
these rings later passes through the SAMBA-east array during the mooring time
period; there were also as four anticyclonic eddies generated by the splitting of one of
the rings. The Agulhas ring statistics (diameter of 170±86 km with
translation speeds of 11±6 km day-1) are in good agreement with
previous estimates (diameter of 200–400 km – e.g., Arhan et al., 1999;
translation speed of 2.9–7.3 km day-1 – e.g., Olson and Evans, 1986;
Byrne et al., 1995; Goni et al., 1997; Schouten et al., 2000). The
comparisons reveal that satellite data provide a reasonable description of
the upper-ocean circulation along the continental slope at our mooring
locations, except near the shelf break (i.e., along and inshore of the
1200 m isobath).
Analyses of both satellite and mooring data show that the eastern mooring
array is strongly affected by the intense regional mesoscale variability.
Previous studies (Arhan et al., 1999, 2011; Schouten et al., 2000; Boebel et
al., 2003) have shown that Agulhas rings coexist with cold-core (cyclonic)
eddies, which can contribute directly to the input of Indian water in the
Atlantic (Lutjeharms et al., 2003; Arhan et al., 2011; Capuano et al., 2018).
The resulting interaction of cyclonic and anticyclonic eddies can be also
responsible for the extraction of warm Agulhas water filaments (Lutjeharms
and Cooper, 1996; Whittle et al., 2008). These filaments may not provide more
than 15 % of the total mass flux of Indian Ocean waters into the South
Atlantic (Lutjeharms and Cooper, 1996).
Over the measurement period from September 2014 to December 2015, the slope
moorings (M2 and M3) are shown to be affected essentially by cyclonic eddies
of different origins. Indeed, these moorings are affected by one cyclonic
eddy generated at the Agulhas Bank (C4), one by the splitting of C4 and four
along the South Benguela upwelling front. The offshore mooring (M4) is
affected by the more complex dynamics characterizing the Cape Basin involving
five Agulhas rings and two anticyclonic and cyclonic eddies both generated
along the South Benguela upwelling front. The propagation of two warm surface
filaments has been highlighted. The presence of several of these mesoscale
features induces intense intra-seasonal upper-ocean velocity variations and
water mass exchange across both the shelf and the open ocean and between the
subantarctic and the subtropical frontal zones.
Our study indicates that exchange of water masses across the continental
slope happens through water advection – not only via mesoscale eddies but
also through wide filaments engendered by the interaction among eddies and,
in particular, through the existence of intense dipoles. As illustrated in
previous studies, such filaments can extend well into the thermocline and can
be related to dipole dynamics (de Steur et al., 2004; Baker-Yeboah et
al., 2010b). These wide intrusions cause intense north–northwestward
currents affecting the whole water column. These injections are different
from ordinary filaments, which exhibit a much smaller vertical extension
(around 300 m of the water column). Among the different processes observed
along the SAMBA-east array, the most significant event is the intrusion of
waters of Indian Ocean and subantarctic origin due to the presence of an
intense dipole.
In terms of the number of occurrences of each type of event during the 14
months of the record, we account for two intrusions of waters associated with
the presence of a dipole, five Agulhas rings, six cyclonic eddies, two
anticyclonic eddies and two warm filaments. Our work suggests that it is not
only the advection of water within Agulhas rings or cyclonic eddies that is
important but that dipole intrusions and filaments also have a significant
impact on the total mass, heat and salt fluxes, and therefore, they all need
to be better accounted for.
The presence of eddies, filaments and the interaction of cyclonic and
anticyclonic eddies have also been described in more detail in this study
with our three case studies. Following the Lilly and Rhines (2002) method, an
assessment of coherent eddies, dipoles and filaments in mooring data has been
achieved. This method allowed us to evaluate the eddy parameters, such as the
eddy apparent radius, the direction of the mean flow and the Rossby number,
all essential elements to characterize eddies in single-point measurements.
The estimation of a small Rossby number (∼0.1) associated with these
features reveals that eddies are not highly nonlinear or ageostrophic by this
measure, but the features are nonlinear in the sense that the corresponding
Rossby number is not nil, so some ageostrophic processes are occurring. The
momentum equation is then dominated by the quasi-geostrophic balance between
the pressure gradient and the Coriolis forces and implies that altimetry data
are adequate for investigating the dynamics of the observed mesoscale
features. From altimetry data, these eddies have a maximal azimuthal velocity
exceeding their translation speed, confirming that the observed mesoscale
features are nonlinear by the metric of Chelton et al. (2011). This
definition is maybe the most pertinent in the context of this study, since it
determines the ability of the observed mesoscale features to advect a parcel
of trapped fluid as they move
(Flierl, 1981).
During the first two case studies, the typical impact of cyclonic and
anticyclonic eddies causing, respectively, an uplift and downward motion of
isopycnal layers is revealed. For the dipole case study, a downward motion of
isopycnal layers is recorded associated with a vertical movement of the
mooring.
The different properties of each type event (cyclonic eddy vs. anticyclonic
eddy; cold vs. warm dipole intrusions) have been compared between full-depth
moorings and CPIES measurements in density space allowing a better
characterization of the full water-column hydrographic properties and the
opportunity to distinguish the changes in temperature and salinity due to
vertical motion (isopycnal and/or mooring displacement) vs. trapped water
masses. In the upper part of the water column, the presence of Indian water
trapped inside the Agulhas rings or advected within dipoles has been
identified by both data sets. The intrusion of subantarctic water into the
upper water layers due to the dipole dynamics is also highlighted. Associated
with these upper intrusions of Indian and subantarctic waters due to the
presence of dipoles, high-salinity I-AAIW at intermediate depth and NADW at
the deepest level are also illustrated. The presence of intermediate
high-salinity I-AAIW is also evidenced during the period of the cyclonic eddy
crossing.
It has been shown that the trapping depth of rings can reach the seafloor
(van Aken et al., 2003). The analyses of our tall mooring deep SBE MicroCat's
data show that not only Agulhas rings but also water intrusions due to the
presence of dipoles extend to 4400 m of depth, impacting the NADW layers and
even deeper layers.
This study presents the first combination of full-depth moorings and
reconstructed fields from CPIES data combined with a GEM technique in the
upper water column within the eastern South Atlantic region. Here properties
are well resolved by the combination of local sensors and by the GEM
reconstructed fields. The reconstructed fields capture the same changes in
the temperature and salinity variability in the upper water column as do the
local sensors' data when mooring motion is smaller than the isopycnal
displacement. Relatively small differences can be attributed to the limited
vertical sampling resolution in the upper 500 m and the mixing associated
with the mesoscale activity in the Cape Basin. When vertical mooring motion
is larger than the isopycnal displacement, the local sensors sample a large
range of neutral density and can overestimate the anomalies during the event
compared to the GEM reconstructed fields. However, the deep water column
properties remain to be analyzed with local temperature and salinity sensors.
Moreover, the distance between CPIES and full-depth moorings (e.g., 210 km
between M3 and M4) does not allow precise transport estimates – as the
typical velocity decorrelation length scale is smaller than this distance
(∼100 km – e.g., Donohue et al., 2010; Meinen et al., 2017).
Finally, this work presents the first independent observation comparison
between full-depth moorings and CPIES data sets along the SAMBA-east array
that gives some evidence of eastern boundary buoyancy anomalies associated
with migrating eddies. It also highlights the need to continuously sample the
full water depth as inter-basin exchange occurs intermittently and affects
the whole water column. Future investigations with longer time series at
these existing sites will lead to a better understanding of the eastern
boundary current variability and Indo-Atlantic exchanges. The impact of each
isolated mesoscale eddy will not be adequately resolved at this scale, but
the global eddy thickness flux anomalies can be improved. The CPIES records
used in combination with the moored instruments in the western part of the
SAMBA array will improve of our understanding of the strength and variability
in the AMOC.
Data used in this study are freely available via several
web pages. Users can access the daily travel time CPIES data on the SAMOC
initiative web page (www.aoml.noaa.gov/phod/SAMOC_international/, last
access: 21 August 2018). Data from the hydrographic casts and the full-depth
moorings are being uploaded to the “Marine Information Management System”
(data.ocean.gov.za/, last access: 21 August 2018). Data from the South
African DEA are also available from Tarron Lamont (tarron.lamont@gmail.com).
Argo data were collected and made freely available by the Coriolis project
and programmes that contribute to it (www.coriolis.eu.org, last access:
21 August 2018). CTD profiles were obtained from the World Ocean Database
(www.nodc.noaa.gov/OC5/indprod.html, last access: 21 August 2018, Boyer
et al., 2013). The altimeter products were produced by Ssalto/Duacs and
distributed by Aviso, with support from CNES
(www.aviso.oceanobs.com/duacs/, last access: 21 August 2018). Odyssea
SST data were produced by the Medspiration project and were obtained from the
Centre de Recherche et d'Exploitation Satellitaire (CERSAT,
cersat.ifremer.fr/data/, last access: 21 August 2018), at IFREMER,
Plouzane (France). The Mesoscale Eddy Trajectory Atlas products were produced recently by Centre
National d'Etudes Spatiales (CNES) Collecte Localisations Satellites (CLS) in
the Data Unification and Altimeter Combination System (DUACS) system and
distributed by Archiving, Validation and Interpretation of Satellite
Oceanographic Data (AVISO)(www.aviso.altimetry.fr/,
last access: 21 August 2018) with support from CNES, in collaboration with
Oregon State University with support from NASA.
MK took the lead in analyzing the data and writing the manuscript.
TL, SS, IA, and MR designed and planned the array, with assistance from TT
and MvdB. RL developed the eddy detection method and performed the eddy
statistics. All authors helped shape the research, and were involved in the
analysis and in preparing the manuscript.
The authors declare that they have no conflict of
interest.
Acknowledgements
The authors would like to express their great appreciation to the captain,
officers and crew of the research vessels which have supported this program
to date, including the South African research vessels the RV Algoa
and the RV SA Agulhas II. We are warmly grateful to the technical
staff who worked on the preparation, deployment and the recovery of the
instruments. And our thanks to those who have helped coordinate these
challenging international cruise collaborations. The authors acknowledge the
support of grants from the NRF/SANAP–SAMOC-SA programme. Marion Kersalé
acknowledges support from a NRF grant via a South African post-doctoral
fellowship. Marion Kersalé's work on this study was carried out in part
under the auspices of the Cooperative Institute for Marine and Atmospheric
Studies (CIMAS), a Cooperative Institute of the University of Miami and the
National Oceanic and Atmospheric Administration (NOAA), cooperative agreement
NA10OAR4320143. Marion Kersalé also acknowledges support from the NOAA
Atlantic Oceanographic and Meteorological Laboratory. This work was also
supported by the European Union's Horizon 2020 research and innovation
programme under grant agreement no. 633211 (AtlantOS) and the 11-ANR-56-004
grant for Sabrina Speich. The authors thank Renellys C. Perez, Chris S.
Meinen and Jonathan Lilly for precious help, comments and useful discussions.
Finally, we thank the editor and the three reviewers of this paper for their
constructive comments.
Edited by: Piers Chapman
Reviewed by: Johannes Karstensen and two anonymous referees
ReferencesAhlnäs, K., Royer, T. C., and George, T. H.: Multiple dipole eddies in
the Alaska Coastal Current detected with Landsat thematic mapper data,
J. Geophys. Res., 92, 13041–13047, 10.1029/JC092iC12p13041, 1987.
Ansorge, I. J., Speich, S., Lutjeharms, J. R. E., Goni, G. J., Rautenbach, C.
D. W., Froneman, P. W., Rouault, M., and Garzoli, S.: Monitoring the oceanic
flow between Africa and Antarctica: report of the first GoodHope cruise:
research in action, S. Afr. J. Sci., 101, 29–35, 2005.Ansorge, I. J., Baringer, M. O., Campos, E. J. D., Dong, S., Fine, R. A.,
Garzoli, S. L., Goni, G., Meinen, C. S., Perez, R. C., Piola, A. R., Roberts,
M. J., Speich, S., Sprintall, J., Terre, T., and Van den Berg, M. A.:
Basin-Wide Oceanographic Array Bridges the South Atlantic, Eos T. Am.
Geophys. Un., 95, 53–54, 10.1002/2014EO060001, 2014.Arhan, M., Mercier, H., and Lutjeharms, J. R. E.: The disparate evolution of
three Agulhas rings, J. Geophys. Res., 104, 20987–21005,
10.1029/1998JC900047, 1999.Arhan, M., Speich, S., Messager, C., Dencausse, G., Fine, R., and Boye, M.:
Anticyclonic and cyclonic eddies of subtropical origin in the subantarctic
zone south of Africa, J. Geophys. Res., 116, C11004,
10.1029/2011JC007140, 2011.Baker-Yeboah, S., Flierl, G. R., Sutyrin, G. G., and Zhang, Y.:
Transformation of an Agulhas eddy near the continental slope, Ocean Sci., 6,
143–159, 10.5194/os-6-143-2010, 2010a.Baker-Yeboah, S., Byrne, D. A., and Watts, D. R.: Observations of mesoscale
eddies in the South Atlantic Cape Basin: Baroclinic and deep barotropic eddy
variability, J. Geophys. Res., 115, C12069, 10.1029/2010JC006236,
2010b.Bang, N. D.: Characteristics of an intense ocean frontal system in the upwell
regime west of Cape Town, Tellus, 25, 256–265,
10.3402/tellusa.v25i3.9659, 1973.Beal, L. M., De Ruijter, W. P., Biastoch, A., and Zahn, R.: On the role of
the Agulhas system in ocean circulation and climate, Nature, 472, 429–436,
10.1038/nature09983, 2011.Beal, L. M., Elipot, S., Houk, A., and Leber, G. M.: Capturing the transport
variability of a western boundary jet: Results from the Agulhas Current
Time-Series Experiment (ACT), J. Phys. Oceanogr., 45, 1302–1324,
10.1175/JPO-D-14-0119.1, 2015.Biastoch, A., Böning, C. W., and Lutjeharms, J. R. E.: Agulhas Leakage
dynamics affects decadal variability in Atlantic overturning circulation,
Nature, 456, 489–492, 10.1038/nature07426, 2008.Biastoch, A., Durgadoo, J. V., Morrison, A. K., Van Sebille, E., Weijer, W.,
and Griffies, S. M.: Atlantic multi-decadal oscillation covaries with Agulhas
leakage, Nature, 6, 10082, 10.1038/ncomms10082, 2015.Boebel, O., Lutjeharms, J., Schmid, C., Zenk, W., Rossby, T., and Barron, C.:
The Cape Cauldron: a regime of turbulent inter-ocean exchange. Deep-Sea Res.
Pt. II, 50, 57–86, 10.1016/S0967-0645(02)00379-X, 2003.Boyd, A. J., Taunton-Clark, J., and Oberholster, G. P. J.: Spatial features
of the near-surface and midwater circulation patterns off western and
southern South Africa and their role in the life histories of various
commercially fished species, Afr. J. Mar. Sci., 12, 189–206,
10.2989/02577619209504702, 1992.Boyer, T. P., Antonov, J. I., Baranova, O. K., Coleman, C., Garcia, H. E.,
Grodsky, A., Johnson, D. R., Locarnini, R. A., Mishonov, A. V., O'Brien, T
.D., Paver, C. R., Reagan, J. R., Seidov, D., Smolyar, I. V., and Zweng, M.
M.: World Ocean Database 2013, NOAA Atlas NESDIS 72, edited by: Levitus, S.
and Mishonov, A., Silver Spring, MD, 209 pp., 10.7289/V5NZ85MT, 2013.Byrne, D. A., Gordon, A. L., and Haxby, W. F.: Agulhas eddies: A synoptic
view using Geosat ERM data, J. Phys. Oceanogr., 25, 902–917,
10.1175/1520-0485(1995)025<0902:AEASVU>2.0.CO;2,
1995.Capuano, T. A., Speich, S., Carton, X., and Blanke, B.: Mesoscale and
Submesoscale Processes in the Southeast Atlantic and Their Impact on the
Regional Thermohaline Structure, J. Geophys. Res.-Oceans, 123, 1937–1961,
10.1002/2017JC013396, 2018.Chaigneau, A., Gizolme, A., and Grados, C.: Mesoscale eddies off Peru in
altimeter records: Identification algorithms and eddy spatio-temporal
patterns, Prog. Oceanogr., 79, 106–119, 10.1016/j.pocean.2008.10.013,
2008.Chaigneau, A., Eldin, G., and Dewitte, B.: Eddy activity in the four major
upwelling systems from satellite altimetry (1992–2007), Prog. Oceanogr., 83,
117–123, 10.1016/j.pocean.2009.07.012, 2009.Chaigneau, A., Le Texier, M., Eldin, G., Grados, C., and Pizarro, O.:
Vertical structure of mesoscale eddies in the eastern South Pacific Ocean: A
composite analysis from altimetry and Argo profiling floats, J. Geophys.
Res., 116, C11025, 10.1029/2011JC007134, 2011.Chelton, D. B., Schlax, M. G., and Samelson, R. M.: Global observations of
nonlinear mesoscale eddies, Prog. Oceanogr., 91, 167–216,
10.1016/j.pocean.2011.01.002, 2011.Chereskin, T. K., Donohue, K. A., Watts, D. R., Tracey, K. L., Firing, Y. L.,
and Cutting, A. L.: Strong bottom currents and cyclogenesis in Drake Passage,
Geophys. Res. Lett., 36, L23602, 10.1029/2009GL040940, 2009.Chidichimo, M. P., Donohue, K. A., Watts, D. R., and Tracey, K. L.:
Baroclinic transport time series of the Antarctic Circumpolar Current
measured in Drake Passage, J. Phys. Oceanogr., 44, 1829–1853,
10.1175/JPO-D-13-071.1, 2014.de Jong, M. F., Bower, A. S., and Furey, H. H.: Two years of observations of
warm-core anticyclones in the Labrador Sea and their seasonal cycle in heat
and salt stratification, J. Phys. Oceanogr., 44, 427–444,
10.1175/JPO-D-13-070.1, 2014.Dencausse, G., Arhan, M., and Speich, S.: Routes of Agulhas rings in the
southeastern Cape Basin, Deep-Sea Res. Pt. I, 57, 1406–1421,
10.1016/j.dsr.2010.07.008, 2010.de Ruijter, W. P. M., Biastoch, A., Drijfhout, S. S., Lutjeharms, J. R. E.,
Matano, R. P., Pichevin, T., van Leeuwen, P. J., and Weijer, W.:
Indian-Atlantic interocean exchange: Dynamics, estimation and impact,
J. Geophys. Res., 104, 20885–20910, 10.1029/1998JC900099, 1999.de Ruijter, W. P. M., van Aken, H. M., Beier, E. J., Lutjeharms, J. R.,
Matano, R. P., and Schouten, M. W.: Eddies and dipoles around South
Madagascar: formation, pathways and large-scale impact, Deep-Sea Res. Pt. I,
51, 383–400, 10.1016/j.dsr.2003.10.011, 2004.de Steur, L., Van Leeuwen, P. J., and Drijfhout, S. S.: Tracer leakage from
modeled Agulhas rings, J. Phys. Oceanogr., 34, 1387–1399,
10.1175/1520-0485(2004)034<1387:TLFMAR>2.0.CO;2,
2004.Donners, J., Drijfhout, S. S., and Hazeleger, W.: Water mass transformation
and subduction in the South Atlantic, J. Phys. Oceanogr., 35, 1841–1860,
10.1175/JPO2782.1, 2005.Donohue, K. A., Watts, D. R., Tracey, K. L., Greene, A. D., and Kennelly, M.:
Mapping circulation in the Kuroshio Extension with an array of current and
pressure recording inverted echo sounders, J. Atmos. Ocean. Tech., 27,
507–527, 10.1175/2009JTECHO686.1, 2010.Donohue, K. A., Tracey, K. L., Watts, D. R., Chidichimo, M. P., and
Chereskin, T. K.: Mean Antarctic Circumpolar Current transport measured in
Drake Passage, Geophys. Res. Lett., 43, 11760–11767,
10.1002/2016GL070319, 2016.Drijfhout, S. S., Weber, S. L., and van der Swaluw, E.: The stability of the
MOC as diagnosed from model projections for pre-industrial, present and
future climates, Clim. Dynam., 37, 1575–1586,
10.1007/s00382-010-0930-z, 2011.Duncombe Rae, C. M.: Agulhas retroflection rings in the South Atlantic Ocean:
an overview, Afr. J. Mar. Sci., 11, 327–344,
10.2989/025776191784287574, 1991.Duncombe Rae, C. M.: A demonstration of the hydrographic partition of the
Benguela upwelling ecosystem at 26∘40′ S, Afr. J. Mar. Sci., 27,
617–628, 10.2989/18142320509504122, 2005.Duncombe Rae, C. D., Shillington, F. A., Agenbag, J. J., Taunton-Clark, J.,
and Gründlingh, M. L.: An Agulhas ring in the South Atlantic Ocean and
its interaction with the Benguela upwelling frontal system, Deep-Sea Res.
Pt. I, 39, 2009–2027, 10.1016/0198-0149(92)90011-H, 1992.Duncombe Rae, C. M., Garzoli, S. L., and Gordon, A. L.: The eddy field of the
southeast Atlantic Ocean: A statistical census from the Benguela Sources and
Transports Project, J. Geophys. Res., 101, 11949–11964,
10.1029/95JC03360, 1996.Elipot, S. and Beal, L. M.: Observed Agulhas Current sensitivity to
interannual and long-term trend atmospheric forcings, J. Climate, 31,
3077–3098, 10.1175/JCLI-D-17-0597.1, 2018.Faghmous, J. H., Frenger, I., Yao, Y., Warmka, R., Lindell, A., and Kumar,
V.: A daily global mesoscale ocean eddy dataset from satellite altimetry,
Scientific Data, 2, 150028, 10.1038/sdata.2015.28, 2015.Flierl, G. R.: Particle motions in large-amplitude wave fields, Geophys.
Astro. Fluid, 18, 39–74, 10.1080/03091928108208773, 1981.Garzoli, S. L. and Gordon, A. L.: Origins and variability of the Benguela
Current, J. Geophys. Res., 101, 897–906, 10.1029/95JC03221, 1996.Garzoli, S. L. and Matano, R.: The South Atlantic and the Atlantic meridional
overturning circulation, Deep-Sea Res. Pt. II, 58, 1837–1847,
10.1016/j.dsr2.2010.10.063, 2011.Gladyshev, S., Arhan, M., Sokov, A., and Speich, S.: A hydrographic section
from South Africa to the southern limit of the Antarctic Circumpolar Current
at the Greenwich meridian, Deep-Sea Res. Pt. I, 55, 1284–1303,
10.1016/j.dsr.2008.05.009, 2008.Goni, G. J., Garzoli, S. L., Roubicek, A. J., Olson, D. B., and Brown, O. B.:
Agulhas ring dynamics from TOPEX/POSEIDON satellite altimeter data, J. Mar.
Res., 55, 861–883, 10.1357/0022240973224175, 1997.Gordon, A. L.: Indian-Atlantic transfer of thermocline water at the Agulhas
Retroflection, Science, 227, 1030–1033, 10.1126/science.227.4690.1030,
1985.Gordon, A. L.: Interocean exchange of thermocline water, J. Geophys. Res.,
91, 5037–5046, 10.1029/JC091iC04p05037, 1986.Gordon, A. L. and Haxby, W. F.: Agulhas eddies invade the South Atlantic:
Evidence from Geosat altimeter and shipboard conductivity-temperature-depth
survey, J. Geophys. Res., 95, 3117–3125, 10.1029/JC095iC03p03117,
1990.Gordon, A. L., Weiss, R. F., Smethie, W. M., and Warner, M. J.: Thermocline
and intermediate water communication between the South Atlantic and Indian
Oceans, J. Geophys. Res., 97, 7223–7240, 10.1029/92JC00485, 1992.Hall, C. and Lutjeharms, J. R. E.: Cyclonic eddies identified in the Cape
Basin of the South Atlantic Ocean, J. Marine Syst., 85 1–10,
10.1016/j.jmarsys.2010.10.003, 2011.Heywood, K. J. and King, B. A.: Water masses and baroclinic transports in the
South Atlantic and Southern oceans, J. Mar. Res., 60, 639–676,
10.1357/002224002762688687, 2002.
Hogg, N. G. and Stommel, H. M.: The heton, an elementary interaction between
discrete baroclinic geostrophic vortices, and its implications concerning
eddy heat-flow, P. Roy. Soc. Lond. A Mat., 397, 1–20, 1985.Hooker, S. B. and Brown, J. W.: Warm core ring dynamics derived from
satellite imagery, J. Geophys. Res., 99, 25181–25194,
10.1029/94JC02171, 1994.Hooker, S. B., Brown, J. W., Kirwan, A. D., Lindemann, G. J., and Mied, R.
P.: Kinematics of a warm-core dipole ring, J. Geophys. Res., 100,
24797–24809, 10.1029/95JC02900, 1995.Hutchinson, K., Swart, S., Meijers, A., Ansorge, I., and Speich, S.:
Decadal-scale thermohaline variability in the Atlantic sector of the Southern
Ocean, J. Geophys. Res., 121, 3171–3189, 10.1002/2015JC011491, 2016.
IOC, SCOR and IAPSO: The international thermodynamic equation of seawater –
2010: Calculation and use of thermodynamic properties, Intergovernmental
Oceanographic Commission (IOC), Manuals and Guides No. 56, UNESCO, 196 pp.,
2010.Kanzow, T., Send, U., Zenk, W., Chave, A. D., and Rhein, M.: Monitoring the
integrated deep meridional flow in the tropical North Atlantic: Long-term
performance of a geostrophic array, Deep-Sea Res. Pt. I, 53, 528–546,
10.1016/j.dsr.2005.12.007, 2006.Lamont, T., Hutchings, L., Van Den Berg, M. A., Goschen, W. S., and Barlow,
R. G.: Hydrographic variability in the St. Helena Bay region of the southern
Benguela ecosystem. J. Geophys. Res., 120, 2920–2944,
10.1002/2014JC010619, 2015.Lilly, J. M. and Rhines, P. B.: Coherent eddies in the Labrador Sea observed
from a mooring, J. Phys. Oceanogr., 32, 585–598,
10.1175/1520-0485(2002)032<0585:CEITLS>2.0.CO;2,
2002.Lilly, J. M., Rhines, P. B., Schott, F., Lavender, K., Lazier, J., Send, U.,
and D'Asaro, E.: Observations of the Labrador Sea eddy field, Prog.
Oceanogr., 59, 75–176, 10.1016/j.pocean.2003.08.013, 2003.
Locarnini, R. A., Mishonov, A. V., Antonov, J. I., Boyer, T. P., Garcia, H.
E., Baranova, O. K., Zweng, M. M. , Paver, C. R., Reagan, J. R., Johnson, D.
R., Hamilton, M., and Seidov, D.: World Ocean Atlas 2013, Volume 1:
Temperature, edited by: Levitus, S. and Mishonov, A., NOAA Atlas NESDIS 73,
40 pp., 2013.Lutjeharms, J. R. E.: The Agulhas Current, Springer Berlin Heidelberg,
329 pp., 10.1007/3-540-37212-1, 2006.Lutjeharms, J. R. E. and Cooper, J.: Interbasin leakage through Agulhas
Current filaments, Deep-Sea Res. Pt. I, 43, 213217–215238,
10.1016/0967-0637(96)00002-7, 1996.Lutjeharms, J. R. E. and Meeuwis, J. M.: The extent and variability of
South-East Atlantic upwelling, Afr. J. Mar. Sci., 5, 51–62,
10.2989/025776187784522621, 1987.
Lutjeharms, J. R. E., Boebel, O., and Rossby, T.: KAPEX: an international
experiment to study deep water movement around southern Africa, S. Afr.
J. Sci., 93, 377–381, 1997.Lutjeharms, J. R. E., Boebel, O., and Rossby, H. T.: Agulhas cyclones,
Deep-Sea Res., Pt. II, 50, 13–34, 10.1016/S0967-0645(02)00378-8, 2003.Matano, R. P. and Beier, E. J.: A kinematic analysis of the Indian/Atlantic
interocean exchange, Deep-Sea Res. Pt. II, 50, 229–249,
10.1016/S0967-0645(02)00395-8, 2003.McDonagh, E. L., Heywood, K. J., and Meredith, M. P.: On the structure,
paths, and fluxes associated with Agulhas rings, J. Geophys. Res., 104,
21007–21020, 10.1029/1998JC900131, 1999.Meinen, C. S and Watts, D. R.: Vertical structure and transport on a transect
across the North Atlantic Current near 42∘ N: Time series and mean,
J. Geophys. Res., 105, 21869–21891, 10.1029/2000JC900097, 2000.Meinen, C. S., Luther, D. S., and Baringer, M. O.: Structure, transport and
potential vorticity of the Gulf Stream at 6∘ W: Revisiting older
data sets with new techniques, Deep-Sea Res. Pt. I, 56, 41–60,
10.1016/j.dsr.2008.07.010, 2009.Meinen, C. S., Piola, A. R., Perez, R. C., and Garzoli, S. L.: Deep Western
Boundary Current transport variability in the South Atlantic: preliminary
results from a pilot array at 34.5∘ S, Ocean Sci., 8, 1041–1054,
10.5194/os-8-1041-2012, 2012.Meinen, C. S., Speich, S., Perez, R. C., Dong, S., Piola, A. R., Garzoli, S.
L., Baringer, M. O., Gladyshev, S., and Campos, E. J. D.: Temporal
variability of the meridional overturning circulation at 34.5∘ S:
Results from two pilot boundary arrays in the South Atlantic, J. Geophys.
Res., 118, 6461–6478, 10.1002/2013JC009228, 2013.Meinen, C. S., Garzoli, S. L., Perez, R. C., Campos, E., Piola, A. R.,
Chidichimo, M. P., Dong, S., and Sato, O. T.: Characteristics and causes of
Deep Western Boundary Current transport variability at 34.5∘ S
during 2009–2014, Ocean Sci., 13, 175–194, 10.5194/os-13-175-2017,
2017.Meinen, C. S., Speich, S., Piola, A. R., Ansorge, I. J., Campos, E. J. D.,
Kersalé , M., Terre, T., Chidichimo, M. P., Lamont, T., Sato, O. T.,
Perez, R. C., Valla, D., van den Berg, M. A., Le Hénaff, M., Dong, S.,
and Garzoli, S. L.: Meridional Overturning Circulation transport variability
at 34.5∘ S during 2009–2017: Baroclinic and barotropic flows and
the dueling influence of the boundaries, Geophys. Res. Lett., 45, 4180–4188,
10.1029/2018GL077408, 2018.
Millot, C. and Taupier-Letage, I.: Circulation in the Mediterranean sea, in:
The Mediterranean Sea, Springer, Berlin, Heidelberg, 323–334, 2005.Nelson, G., Boyd, A. J., Agenbag, J. J., and Duncombe Rae, C. M.: An
upwelling filament north-west of Cape Town, South Africa, Afr. J. Mar. Sci.,
19, 75–88, 10.2989/025776198784126953, 1998.Nencioli, F., Dong, C., Dickey, T., Washburn, L., and McWilliams, J. C.: A
vector geometry–based eddy detection algorithm and its application to a
high-resolution numerical model product and high-frequency radar surface
velocities in the Southern California Bight, J. Atmos. Ocean. Tech., 27,
564–579, 10.1175/2009JTECHO725.1, 2010.Olson, D. B. and Evans, R. H.: Rings of the Agulhas current, Deep-Sea Res.
Pt. I, 33, 27–42, 10.1016/0198-0149(86)90106-8, 1986.Pallàs-Sanz, E. and Viúdez, Á.: Three-dimensional ageostrophic
motion in mesoscale vortex dipoles, J. Phys. Oceanogr., 37, 84–105,
10.1175/JPO2978.1, 2007.Pegliasco, C., Chaigneau, A., and Morrow, R.: Main eddy vertical structures
observed in the four major Eastern Boundary Upwelling Systems, J. Geophys.
Res., 120, 6008–6033, 10.1002/2015JC010950, 2015.Penven, P., Lutjeharms, J. R. E., Marchesiello, P., Roy, C., and Weeks, S.
J.: Generation of cyclonic eddies by the Agulhas Current in the lee of the
Agulhas Bank, Geophys. Res. Lett., 28, 1055–1058,
10.1029/2000GL011760, 2001.Perez, R. C., Garzoli, S. L., Meinen, C. S., and Matano, R. P.: Geostrophic
velocity measurement techniques for the meridional overturning circulation
and meridional heat transport in the South Atlantic, J. Atmos. Ocean. Tech.,
28, 1504–1521, 10.1175/JTECH-D-11-00058.1, 2011.Pujol, M.-I., Faugère, Y., Taburet, G., Dupuy, S., Pelloquin, C., Ablain,
M., and Picot, N.: DUACS DT2014: the new multi-mission altimeter data set
reprocessed over 20 years, Ocean Sci., 12, 1067–1090,
10.5194/os-12-1067-2016, 2016.Richardson, P. L. and Garzoli, S. L.: Characteristics of intermediate water
flow in the Benguela current as measured with RAFOS floats, Deep-Sea Res.
Pt. II, 50, 87–118, 10.1016/S0967-0645(02)00380-6, 2003.Richardson, P. L., Lutjeharms, J. R. E., and Boebel, O.: Introduction to the
“Inter-ocean exchange around southern Africa”, Deep-Sea Res. Pt. I, 50,
1–12, 10.1016/S0967-0645(02)00376-4, 2003.Robinson, A. R. (Ed.): Overview and summary of eddy science, in: Eddies in
marine science, Springer, Berlin, Heidelberg, 3–15,
10.1007/978-3-642-69003-7_1, 1983.Roemmich, D. and Cornuelle, B.: The subtropical mode waters of the South
Pacific Ocean, J. Phys. Oceanogr., 22, 1178–1187,
10.1175/1520-0485(1992)022<1178:TSMWOT>2.0.CO;2,
1992.Roman, R. E. and Lutjeharms, J. R. E.: Red sea intermediate water at the
Agulhas current termination, Deep-Sea Res. Pt. I, 54, 1329–1340,
10.1016/j.dsr.2007.04.009, 2007.Rusciano, E., Speich, S., and Ollitrault, M.: Interocean exchanges and the
spreading of Antarctic Intermediate Water south of Africa, J. Geophys. Res.,
117, C10010, 10.1029/2012JC008266, 2012.
Schiermeier, Q.: Oceans under surveillance, Nature, 497, 167–169, 2013.Schouten, M. W., de Ruijter, W. P. M., van Leeuwen, P. J., and Lutjeharms, J.
R. E.: Translation, decay and splitting of Agulhas rings in the southeastern
Atlantic Ocean, J. Geophys. Res., 105, 21913–21925,
10.1029/1999JC000046, 2000.Shannon, L. V. and Nelson, G.: The Benguela: large scale features and
processes and system variability, in: The South Atlantic, Springer, Berlin,
Heidelberg, 163–210, 10.1007/978-3-642-80353-6_9, 1996.Spall, M. A.: Frontogenesis, subduction, and cross-front exchange at upper
ocean fronts, J. Geophys. Res., 100, 2543–2557, 10.1029/94JC02860,
1995.
Speich, S., Arhan, M., Ansorge, I., Boebel, O., Sokov, A., Gladyshev, S.,
Farbach, E., Byrne, D., Klepikov, A., and Garzoli, S.: GOODHOPE/Southern
Ocean: A study and monitoring of the Indotlantic connections, Mercator
Newsletter, 27, 29–41, October 2007.
Speich, S., Garzoli, S., Piola, A., and the SAMOC community: A monitoring
system for the South Atlantic as a component of the MOC, in: Proceedings of
OceanObs'09: Sustained Ocean Observations and Information for Society
(Annex), Venice, Italy, 21–25 September 2009, edited by: Hall, J., Harrison,
D. E., and Stammer, D., ESA Publication WPP-306, 2010.Sutherland, D. A., Straneo, F., Lentz, S. J., and Saint-Laurent, P.:
Observations of fresh, anticyclonic eddies in the Hudson Strait outflow,
J. Marine Syst., 88, 375–384, 10.1016/j.jmarsys.2010.12.004, 2011.Swart, S., Speich, S., Ansorge, I. J., Goni, G. J., Gladyshev, S., and
Lutjeharms, J. R.: Transport and variability of the Antarctic Circumpolar
Current south of Africa, J. Geophys. Res., 113, C09014,
10.1029/2007JC004223, 2008.Thomson, R. E. and Emery, W. J.: Chapter 3 – Statistical Methods and Error
Handling, in: Data analysis methods in physical oceanography, 3rd edn.,
Elsevier, Boston, 219–311 10.1016/B978-0-12-387782-6.00003-X, 2014.Ursella, L., Kovačević, V., and Gačić, M.: Footprints of
mesoscale eddy passages in the Strait of Otranto (Adriatic Sea), J. Geophys.
Res., 116, C04005, 10.1029/2010JC006633, 2011.Valla, D., Piola, A. R., Meinen, C. S., and Campos, E.: Strong mixing and
recirculation in the northwestern Argentine Basin, J. Geophys. Res.-Oceans,
123, 4624–4648, 10.1029/2018JC013907, 2018.van Aken, H. M., Van Veldhoven, A. K., Veth, C., De Ruijter, W. P. M., Van
Leeuwen, P. J., Drijfhout, S. S., Whittle, C. P., and Rouault, M.:
Observations of a young Agulhas ring, Astrid, during MARE in March 2000,
Deep-Sea Res. Pt. II, 50, 167–195, 10.1016/S0967-0645(02)00383-1,
2003.van Ballegooyen, R. C., Gründlingh, M. L., and Lutjeharms, J. R. E.: Eddy
fluxes of heat and salt from the southwest Indian Ocean into the southeast
Atlantic Ocean: A case study, J. Geophys. Res., 99, 14053–14070,
10.1029/94JC00383, 1994.van den Berg, M.: Cruise report: SAMBA Moorings & Monitoring Line, RS Algoa
Voyage 221, 30 November–6 December 2015, Department of Environmental
Affairs, RSA, Cruise Reports, available at:
http://www.aoml.noaa.gov/phod/SAMOC_international/documents/Cruise_Report_alg221_final.pdf
(last access: 14 August 2018), 2015.van Sebille, E., England, M. H., and Froyland, G.: Origin, dynamics and
evolution of ocean garbage patches from observed surface drifters, Environ.
Res. Lett, 7, 044040, 10.1088/1748-9326/7/4/044040, 2012.Veitch, J., Hermes, J., Lamont, T., Penven, P., and Dufois, F.: Shelf-edge
jet currents in the southern Benguela: A modelling approach, J. Marine Syst.,
10.1016/j.jmarsys.2017.09.003, in press, 2017.Watts, D. R. and Kontoyiannis, H.: Deep-ocean bottom pressure measurement:
Drift removal and performance, J. Atmos. Ocean. Tech., 7, 296–306,
10.1175/1520-0426(1990)007<0296:DOBPMD>2.0.CO;2,
1990.Watts, D. R., Sun, C., and Rintoul, S.: A two-dimensional gravest empirical
mode determined from hydrographic observations in the Subantarctic Front, J.
Phys. Oceanogr., 31, 2186–2209,
10.1175/1520-0485(2001)031<2186:ATDGEM>2.0.CO;2,
2001.
Whittle, C., Lutjeharms, J. R. E., Rae, D., and Shillington, F. A.:
Interaction of Agulhas filaments with mesoscale turbulence: a case study,
S. Afr. J. Sci., 104, 135–139, 2008.
Zweng, M. M, Reagan, J. R., Antonov, J. I., Locarnini, R. A., Mishonov, A.
V., Boyer, T. P., Garcia, H. E., Baranova, O. K., Johnson, D. R., Seidov, D.,
and Biddle, M. M.: World Ocean Atlas 2013, Volume 2: Salinity, edited by:
Levitus, S. and Mishonov, A., NOAA Atlas NESDIS 74, 39 pp., 2013.