The Antarctic Circumpolar Current (ACC) flows eastwards around the Antarctic continent, transporting roughly between 100 and 173 Sv (1 Sv = 106 m3 s-1≈ 109 kg s-1; hereafter Sv will be the unit used) (Orsi et al., 1995; Cunningham et al., 2003; Donohue et al., 2016). Along its path, it connects the Atlantic, Pacific and Indian basins, exchanging heat and freshwater among other properties. Although convergence of net flux estimates has been achieved on basin scales (Ganachaud and Wunsch, 2003), the ACC flow into the Atlantic Ocean is critical to establish the magnitude and pathways of the Southern Ocean contribution to the deep global ocean ventilation.

Peterson and Whitworth III (1989) suggested that the Subantarctic Front (SAF) and the Polar Front (PF), where the major velocity bands of the ACC occur, turn northwestward across the Falkland Plateau to the west of the Maurice Ewing Bank, along the Patagonian continental slope. This was supported by Peterson (1992), who estimated the large contribution of the ACC to the Falkland Current (60–70 Sv), revealing the importance of the overflow of southern waters to the South Atlantic boundary circulation. Peterson and Whitworth III (1989) located the SAF near 53∘ W, as corroborated by Arhan et al. (2002), at a location where the ocean depth is 2000 m. Several studies have later examined the path of the PF around the Maurice Ewing Bank (Trathan et al., 2000) and its branching around 49–50∘ W (Arhan et al., 2002), with a possible meandering of the front according to Naveira Garabato et al. (2002).

The first hydrographic cruise along the Falkland Plateau was carried out in 1999. The ALBATROSS (Antarctic Large-scale Box Analysis and the Role of the Scotia Sea) cruise explored the ACC through the Drake Passage and the Scotia Sea (Fig. 1). The data of this cruise have been used to estimate relative transport, water masses, fluxes and mixing across the plateau (Naveira Garabato et al., 2003) and to provide a detailed explanation of the deep waters in the Scotia Sea (Naveira Garabato et al., 2002). Later on, this section was compared with hydrographic cruises carried out north and south of the Falkland Plateau to achieve a better knowledge of this area (Arhan et al., 2002; Smith et al., 2010). However, it is not until 2010 that this section is repeated.

In this study, the water masses, relative geostrophic velocities and transports across an almost zonal hydrographic section carried out in 2010 along the Falkland Plateau are evaluated. These data, together with the ALBATROSS cruise, are the only high-resolution hydrographic data available in the region. Thus, results from the 2010 cruise are compared with those obtained from the 1999 cruise in the same area (Naveira Garabato et al., 2003), with the objective of assessing possible relative transport and water mass differences between the two surveys. For changes in the relative transport, the position of the fronts and the season in which each cruise took place will be considered. Changes in water masses are decomposed into changes in the θ/S isobaric surfaces and results from the Bindoff and McDougall (1994) model.

The MOC-Austral cruise was carried out in February 2010 onboard the BIO Hespérides. As shown in Fig. 1a, 27 full-depth conductivity, temperature and depth (CTD) stations were occupied across the Falkland Plateau, tracking along the casts previously conducted in between 41 and 57∘ W at the nominal latitude of 51∘ S during the ALBATROSS cruise in April 1999 (Naveira Garabato et al., 2003). With a spatial separation of 30 to 50 km, temperature and salinity profiles were obtained using a SeaBird 911+ CTD with dual conductivity and temperature sensors. The CTD was sent to SeaBird for calibration before the cruise. The temperature sensor has an accuracy of 0.001 ∘C. The conductivity sensors were calibrated on board with bottle sample salinities. To that end, water samples were analyzed on a Guildline AUTOSAL 8400B salinometer with accuracy better than 0.002 for single samples (salinity is expressed in the Practical Salinity Scale, UNESCO, 1981).

Relative geostrophic velocities and mass transports are estimated for both the ALBATROS and MOC-Austral cruises using the sea bottom as the level of no motion. The water column is divided into 18 neutral density layers following the work of Naveira Garabato et al. (2003) (see Table 1). Weddell Sea Deep Water (WSDW) is not found along the plateau; thus, its density layers are not considered here.

Bindoff and McDougall (1994) describe a model to evaluate the temperature and salinity variations in the water column. This model relates the temperature and salinity in both pressure and density changes through the following equation: dψdtz=dψdtγn-dpdtγndψdp, which shows that for a given property (ψ, temperature or salinity), the variations along isobaric levels dψdtz can be described as the sum of changes along isoneutral surfaces dψdtγn and changes due to vertical displacements of the density surfaces, referred to as heaving dpdtγndψdp. This allows the comparison between the isobaric changes and the sum of the two decomposed components, which represent the variations of the water masses (warming and freshening) and the heaving. To apply this methodology, temperature and salinity are interpolated onto a grid with a pressure interval of 20 dbar (from 10 to 3500 dbar) and the following neutral density (kg m-3) values: from 26 to 27.6 each 0.02, from 27.7 to 28 each 0.01 and from 28.005 to 28.5 each 0.005. This vector is selected to properly represent the different structures found in the water column.

In addition, sea surface height (SSH) was downloaded from AVISO (http://www.aviso.oceanobs.com/, Dibarboure et al., 2015). Data between 10 and 20 February 2010 are used for the analysis of the MOC-Austral cruise results. Data of February and April from 1993 to 2016 are used to estimate the average seasonal geostrophic transport in each month.

The decadal differences in the Falkland Plateau are studied from full-depth hydrographic data collected during the ALBATROSS (April 1999) and MOC-Austral (February 2010) cruises. Water mass changes are explored in terms of changes along neutral surfaces and changes due to vertical displacements of γn surfaces, applying the model proposed by Bindoff and McDougall (1994). Variability in the SAF and PF location and mass transport is inferred from relative geostrophic velocities estimated by using the sea bottom as the level of no motion.

The SASW stratum presents a wider range of salinities and temperatures in 1999 than in 2010, as shown in the θ/S diagram. In spite of this, the θ/S isobaric changes show an increase in surface temperatures and salinities matching the Bindoff and McDougall (1994) model's result for changes along neutral surfaces. The model also exhibits shoaling of the isopycnals. The most plausible source of these differences is the fact that the hydrographic cruises took place in different seasons (ALBATROSS in austral fall and MOC-Austral in austral summer). Hence, the seasonal thermocline has probably changed its depth due to the different seasonal heating and precipitation. The scarce in situ observations in the area do not allow any further conclusions.

SAMW expands over a larger depth range and presents a wider range of salinities in 1999 than in 2010 (Figs. 2 and 4). In contrast, the θ/S diagram and isobaric changes for the AAIW/AASW stratum show a decrease in temperature and salinity in 2010 when the AAIW/AASW occupies a higher depth range (Figs. 2 and 4). As both Bindoff and McDougall (1994) model estimations (with and without frontal zones) agree in that the changes in the intermediate strata are due to the displacement of γn surfaces, some changes are likely to have occurred between 1999 and 2010 in the Falkland Plateau. The Bindoff and McDougall (1994) model reveals a deepening of the isoneutrals at these levels, where the AAIW/AASW stratum occupies a higher depth range than SAMW. An explanation for changes in those strata can be found in Naveira Garabato et al. (2009). Figure 11a shows the mean wind stress of the winters in the period 1998–2010. This figure is analogous to Fig. 11a of Naveira Garabato et al. (2009). In the climatological mean a continuous wind stress magnitude spreads west from South America (Fig. 11a). Figure 11b and c exhibits the previous winter anomalies to the ALBATROSS and MOC-Austral cruises, respectively. These anomalies look very different between themselves. Figure 11b shows a large eastward (positive) wind stress anomaly in the South Pacific. Naveira Garabato et al. (2009) suggest that this structure causes a shift in the SAMW formation area. This matches with the changes observed in Figs. 2 and 4, where the SAMW stratum area is reduced. It also agrees with the isobaric changes reported, a decrease in temperature of 0.8 ∘C and an increase in salinity of 0.02 from 1999 to 2010.

Naveira Garabato et al. (2009) also reported that the 1998 wind stress anomaly pattern shown in Fig. 11b generates a shutdown of the AAIW formation. Due to this, a minimum of temperatures (< 1.2 ∘C) and salinities (ca. 34) can be observed only for the MOC-Austral cruise in Figs. 1b and 2b. The shutdown of the AAIW formation in 1998 is responsible for the observed changes from 1999 to 2010 at this stratum. Across the decade, the AAIW/AASW stratum increases the spanning area at intermediate layers and suffers a decrease of 0.6 ∘C in temperature and of 0.07 in salinity, which is accompanied by a deepening of the isoneutrals. Wind-driven changes in the ACC isobaric surfaces were also observed in Böning et al. (2008), where a deepening of the isopycnals of 27.2 and 27.4 kg m-3 is described. The reported decrease of 0.07 in salinity agrees with the decadal trend of the ACC at 300–500 dbar observed in Böning et al. (2008), shown in their Fig. 4. In contrast, they find an increase in temperature at the same layer, probably due to the contribution of other intermediate waters to the ACC.

The SAF and PF undergo some displacements and variations in intensity between 1999 and 2010. The SAF in 1999 is observed at 52.2–54.8∘ W with a relative mass transport of 32.6 Sv and, while it is wider in 2010, reaching the Falkland Islands, it weakens to roughly half of the transport (17.9 Sv). The SAF is the main path for the northward flow of SASW, SAMW and AAIW/AASW into the Atlantic Basin. The PF also contributes to this northward flow, being important for the UCDW and LCDW. The PF in 1999 is located in the longitudinal range 49.8–44.4∘ W carrying 24.9 Sv, while in 2010 it narrows, centering on 49.9–49∘ W and strengthening to 37.3 Sv. The PF in 2010 carries the highest relative northward transport of the study area, but nearly 8 Sv of it recirculate back southward, as seen in the SSH image. This meandering of the PF was also observed in Naveira Garabato et al. (2002). The average of the AVISO climatological seasonal transport is non-significantly different between February and April. Thus, the observed changes in transport are due to interannual variability.

The abrupt decrease in the SAF to almost half of its transport between both decades has it source in the different wind patterns observed in Fig. 11 and described in Naveira Garabato et al. (2009). The changes in these westerly winds that circle Antarctica are driven by the Southern Annular Mode (SAM). Combes and Matano (2014) analyze the 1990 to 2012 trend of the SAM and its relation to the Falkland Current (which is nourished from the SAF). In their study they report a decrease in the Drake Passage transports and in the Falkland Current after 1999–2000 as a consequence of a weakening of the westerly winds.

To conclude, a seasonal change in the thermocline is observed in the surface layer. Changes in the westerly winds driven by the increasing trend of the SAM (Combes and Matano, 2014; Naveira Garabato et al., 2009) have an important effect on the water masses and transports observable in the Falkland Plateau. The intermediate water masses of the study area seem to be very sensitive to the wind conditions existing in their formation area. Hence, in 2010 an increase (decrease) in the AAIW/AASW (SAMW) stratum is observed together with a cooling, freshening and deepening of the isopycnals at this level. The CDW layers do not exhibit any significant change in the water mass properties, being the most homogenous water mass. However, the LCDW exhibited a temperature- and salinity-driven deepening of the isopycnals from 1999 to 2010. The net transport is 9.2 Sv weaker in 2010 than in 1999 and is mainly explained by a decrease in the transport of the SAF. Fronts change their width and strength between cruises, the SAF/PF being wider/thinner in 2010 and weaker/stronger than in 1999. Changes in the water masses, position and transport contribution of the SAF and PF to the north directly affect the Brazil–Malvinas Confluence Zone, which moves to the south when the SAF weakens (Combes and Matano, 2014). This strong frontal zone is critical for the Southern Hemisphere meridional overturning, is where the deep western boundary separates from the Argentinian basin, and is a high-energy area that contributes to the active water mass transformation (Mason et al., 2017).