1. Introduction
The Antarctic continental slope is a key region for the global climate as it constitutes a natural barrier to the flow between the open ocean and the Antarctic continental shelf (Thompson et al. 2018). Water mass exchanges across the continental slope are critical for the global overturning circulation (Jacobs 2004) and the mass balance of the Antarctic ice sheet, with consequences for the global sea level (DeConto and Pollard 2016). The large topographic gradient across the continental slope, however, steers the flow mostly along the continental slope, resulting in a counterclockwise flowing current known as the Antarctic Slope Current (ASC) (Whitworth et al. 1998). The spatial and temporal variability of the ASC is therefore an important aspect in understanding the climate system.
Despite its global significance, our knowledge of the ASC and its variability is constrained, due to sparse observations of the high-latitude Southern Ocean (Thompson et al. 2018; Pauthenet et al. 2021) and to the high spatial resolution required to resolve the processes on the Antarctic continental slope in numerical models (Nøst et al. 2011; Dinniman et al. 2012; St-Laurent et al. 2013; Hattermann et al. 2014; Stewart and Thompson 2015). Jacobs (1991) was the first to describe the ASC, by gathering the available observations at the time, to be a circumpolar feature that closely follows the continental slope and connects the different ocean basins. Since Jacobs’ (1991) pioneering study, regional observational programs have contributed to the advancement in the understanding of the ASC and provided evidence for a variable circulation, both in space and time. The first circumpolar description of the ASC based on a single dataset stems from Armitage et al. (2018), who use satellite-based sea surface height observations to provide an estimate of the ASC’s geostrophic component. However, the data product’s spatial resolution of approximately 50 km is not adequate to fully resolve the flow structure. Only a couple of modeling studies with a circumpolar domain have focused on the ASC (Mathiot et al. 2011; Stewart et al. 2019), but none of these studies have investigated the spatial and subannual variability of the ASC in an eddying ocean model with dense shelf water formation on the continental shelf. Thus, most existing knowledge of the ASC along the continental slope has been provided only from regional studies.
The available information from observations and models on the ASC along the continental slope can be summarized as follows. The drivers of the ASC are proposed to include (i) easterly winds that result in onshore Ekman transport, which in turn deepens the isopycnals near the continental slope and raise the sea level on the continental shelf (Gill 1973); (ii) a cross-slope density gradient due to, for example, low densities on the continental shelf from basal melting (Thompson et al. 2020); and (iii) tides that can drive a mean flow via a net redistribution of volume onto the continental shelf, resulting in cross-slope surface pressure gradients (Flexas et al. 2015; Stewart et al. 2019). The ASC is strongest in East Antarctica and is weakest, or even reverses direction, in West Antarctica where the Antarctic Peninsula impinges farther north into the Southern Ocean (Thompson et al. 2018). The ASC has a surface expression around the continent, but a bottom intensification of the flow field exists at locations downstream of dense shelf water formation regions (Heywood et al. 1998; Fukamachi et al. 2000; Gordon et al. 2009; Chavanne et al. 2010; Hirano et al. 2015; Azaneu et al. 2017). A bottom-intensified westward flow results from the overflow of the dense waters forming dense plumes that are steered along the continental slope following geostrophy and potential vorticity conservation (Baines 2009). A temporal reversal of the current direction with depth occurs in the Amundsen Sea, the eastern Weddell Sea, and near the Totten Ice Shelf in East Antarctica and has been linked to elevated onshore heat transport (Heywood et al. 1998; Núñez-Riboni and Fahrbach 2009; Chavanne et al. 2010; Walker et al. 2013; Silvano et al. 2019). Modeling work (Stern et al. 2015) and observations from the western Weddell Sea (Azaneu et al. 2017) and East Antarctica (Peña-Molino et al. 2016) further suggest that the ASC consists of multiple jets that drift from the shelf break toward the deep ocean. Last, seasonal variability of the ASC, with larger velocities in autumn and winter, has been related to changes in the winds (Núñez-Riboni and Fahrbach 2009; Peña-Molino et al. 2016; Armitage et al. 2018).
The description of the ASC along the Antarctic continental slope is patchy and lacks a circumpolar overview. Our study is motivated by work on the Antarctic Slope Front (i.e., the density structure over the continental slope), which is associated with the ASC (i.e., the velocity structure over the continental slope) and has been studied more extensively. In particular, Thompson et al. (2018), based on observations, and Moorman et al. (2020), based on a model simulation, categorize the continental shelf into distinct frontal regimes: the fresh shelf, the dense shelf, the warm shelf, and in the case of Moorman et al. (2020) an additional cool shelf regime. Such a classification allows the comparison of different regions and the identification of the spatially varying leading-order dynamics.
The fresh shelf is typically found in East Antarctica and occupies stretches of the continental slope where isopycnals slope downward in an onshore direction due to poleward wind-induced Ekman transport. The isopycnals typically intersect with the bathymetry on the continental slope, preventing an along-isopycnal pathway for cross-slope exchange and creating a gradient in water mass properties across the continental shelf break. The dense shelf occurs when the shelf waters are dense enough to raise the middepth isopycnals near the continental shelf break during overflow events. The isopycnals have their deepest point near the continental slope and shallow toward the continent and the open ocean and thereby allow an isopycnal connection of the water masses across the shelf break. The warm shelf is most prominent in West Antarctica and has a weak frontal structure as the winds are relatively weak and no dense shelf water forms. The shelf waters are warmer compared to other parts of the continental shelf as heat can be exchanged with the offshore ocean and is not lost to the atmosphere. The cool shelf develops in the eastern Ross Sea and has a weak frontal structure; however, different from the warm shelf, it maintains a cross-shelf temperature gradient.
The aim of this study is to provide a comprehensive description of the flow field and its variability along the continental slope. Following the example of the ASF classification, we introduce regimes for the ASC using velocity data from a 0.1° ocean–sea ice model. The model has been shown to achieve good results in terms of simulating the water masses on and near the continental shelf including the formation and export pathways of dense shelf water (Morrison et al. 2020; Moorman et al. 2020). Our intention is to present a large-scale overview of the ASC and its spatial and subannual variability. We show that the spatial distributions of the ASC and ASF regimes align well and we describe characteristics of each ASC regime such as the relationship between the surface, bottom, and barotropic velocity as well as the seasonal variability.
2. Methods and model validation
a. The ocean–sea ice model
This study makes use of the global ocean–sea ice model ACCESS-OM2-01 (for a detailed description see Kiss et al. 2020) with 0.1° horizontal resolution and 75 vertical levels on a z* grid. The ocean model component is MOM version 5.1 (Griffies 2012) and the sea ice model component is CICE version 5.1.2 (Hunke et al. 2015). The atmospheric forcing is prescribed by the JRA55-do v1.3 product (Tsujino et al. 2018), which has a spatial resolution of 55 km and a 3-hourly temporal resolution. The model simulation we analyze has a repeat year forcing from 1 May 1990 to 30 April 1991, a time that is characterized by neutral conditions of the dominant climate variability modes (Stewart et al. 2020). The model is spun up for 180 years and we use 10 years (model years 180–189) of monthly averaged output to look at monthly climatologies or to calculate annual averages and 1 year (model year 186) of daily averaged output to investigate higher temporal variability.
The bathymetry product is based on the GEBCO 2014 30arcsec grid version 20150318a (GEBCO 2014) and has been adapted for the model grid [see Kiss et al. (2020) for a detailed description]. The ocean is bounded by a vertical wall at the location of ice shelf fronts as the ocean model component does not support ice shelf cavities. The freshwater forcing from basal melting and iceberg calving is applied along the coastline at the ocean surface. Estimates of the net freshwater magnitude and the spatial variability for Antarctica in the JRA55-do v1.3 product come from Depoorter et al. (2013), who provide a climatological mean.
b. Comparison of simulated velocities to observations
In this section we provide a comparison of the modeled velocity on the Antarctic continental slope with observations to evaluate the suitability of ACCESS-OM2-01 for the description of the ASC’s spatial and temporal variability. We compare the model output with a mooring transect across the continental slope near −246.7°E (113.3°E) in East Antarctica (Fig. 1). The exact orientation of the transect differs slightly between the observations and the model, which, together with two different bathymetry products (deepest available observations versus model bathymetry), results in the differences in the transect geometry. The modeled transect is along a longitude (i.e., along the model grid) instead of closely following the observed transect, which allows us to calculate the transport across the transect. Further note that the modeled transect extends farther toward the continental shelf. The observational product was published in Peña-Molino et al. (2016); the dataset is hereafter referred to as PM2016 and is to our knowledge the only available long-term mooring transect of the Antarctic continental slope. Five moorings were deployed over a 17-month time period between January 2010 and May 2011. The advantage of the PM2016 dataset is that it provides statistics on both the ASC mean state and its variability in time, with the caveat that the moorings only cover the ocean below 500-m depth to protect them from passing icebergs. Data for the upper 500-m water column are taken from the recovery cruise and therefore do not include the full time series.
(a) Modeled and (b) observed [data replotted from Peña-Molino et al. (2016)] zonal velocity averaged over 17 months (from January to May the following year) for a cross-slope transect along −246.7°E (113.3°E) in East Antarctica. Negative velocities are westward and indicate a flow into the page. The model transect extends farther south, as indicated by the white dashed vertical line, to capture more of the surface-intensified westward current located farther onshore. The observed transect below 500 m stems from a moored array with five moorings as indicated by the triangles and vertical black lines. The data of the upper 500 m is a geostrophic estimate based on hydrographic measurements from the mooring recovery cruise and represents a snapshot in time. Bathymetry is shown in gray and depicts the model bathymetry in (a) and the depth of the deepest observation in (b). (c),(d) Daily time series of maximum cumulative westward transport, summed from the coast northward. The black and orange curves show the transport below 500 m, and the blue line indicates the full transport from surface to bottom. The dashed lines are the time averaged transport, which is also given in the legend together with period of the dominant mode of variability.
Citation: Journal of Physical Oceanography 52, 3; 10.1175/JPO-D-21-0143.1
(a) Modeled and (b) observed [data replotted from Peña-Molino et al. (2016)] zonal velocity averaged over 17 months (from January to May the following year) for a cross-slope transect along −246.7°E (113.3°E) in East Antarctica. Negative velocities are westward and indicate a flow into the page. The model transect extends farther south, as indicated by the white dashed vertical line, to capture more of the surface-intensified westward current located farther onshore. The observed transect below 500 m stems from a moored array with five moorings as indicated by the triangles and vertical black lines. The data of the upper 500 m is a geostrophic estimate based on hydrographic measurements from the mooring recovery cruise and represents a snapshot in time. Bathymetry is shown in gray and depicts the model bathymetry in (a) and the depth of the deepest observation in (b). (c),(d) Daily time series of maximum cumulative westward transport, summed from the coast northward. The black and orange curves show the transport below 500 m, and the blue line indicates the full transport from surface to bottom. The dashed lines are the time averaged transport, which is also given in the legend together with period of the dominant mode of variability.
Citation: Journal of Physical Oceanography 52, 3; 10.1175/JPO-D-21-0143.1
(a) Modeled and (b) observed [data replotted from Peña-Molino et al. (2016)] zonal velocity averaged over 17 months (from January to May the following year) for a cross-slope transect along −246.7°E (113.3°E) in East Antarctica. Negative velocities are westward and indicate a flow into the page. The model transect extends farther south, as indicated by the white dashed vertical line, to capture more of the surface-intensified westward current located farther onshore. The observed transect below 500 m stems from a moored array with five moorings as indicated by the triangles and vertical black lines. The data of the upper 500 m is a geostrophic estimate based on hydrographic measurements from the mooring recovery cruise and represents a snapshot in time. Bathymetry is shown in gray and depicts the model bathymetry in (a) and the depth of the deepest observation in (b). (c),(d) Daily time series of maximum cumulative westward transport, summed from the coast northward. The black and orange curves show the transport below 500 m, and the blue line indicates the full transport from surface to bottom. The dashed lines are the time averaged transport, which is also given in the legend together with period of the dominant mode of variability.
Citation: Journal of Physical Oceanography 52, 3; 10.1175/JPO-D-21-0143.1
Figures 1a and 1b show the modeled and observed time mean, respectively, of the zonal velocity field for the cross-slope transect. The overall pattern is similar between the model and the observations with a much more fine-scale structure in the model data due to the higher spatial resolution (velocity observations with a higher spatial resolution obtained during the mooring recovery cruise confirm the existence of fine-scale structure). The main features are a surface-intensified westward flow at the upper continental slope, a bottom-intensified westward flow with maximum velocities at the seafloor, and a surface-intensified eastward flow farther offshore, typical for an Antarctic Circumpolar Current front that reaches all the way to the bottom. The modeled bottom-intensified westward flow is located farther down the slope and not immediately below the surface-intensified westward ASC, as in the observations.
We further compare the maximum cumulative westward transport, summed from the coast northward, for the water column below 500 m (Figs. 1c,d). The mean transport matches well with 19.0 Sv (1 Sv ≡ 106 m3 s−1) in the model and 20.4 Sv in the observations. (Including the upper 500 m for the modeled transport results in a slightly larger value of 21.1 Sv because of the surface-intensified component of the westward ASC.) The transport is highly variable in time and fluctuates in the model between 0 and 40 Sv with a dominant period of 23 days matching the wind time scale. The PM2016 transport reaches maximum values of 100 Sv for one extreme storm event, but usually does not exceed values larger than approximately 60 Sv. The dominant period of the fluctuations in the observations is 30 days, only slightly larger than in the model.
Overall, the model is able to capture the spatial pattern and to reproduce realistic values for the mean transport and variability at this transect. Due to the large temporal and spatial variability of the ASC, other available observations from the Antarctic margin that are only snapshots in time (of a cross-slope transect) or from a single location (time series from a single mooring) might be misleading and are hence inappropriate for further model validation. The agreement with PM2016, however, encourages us to make use of the model’s large spatial and temporal coverage to study the flow field along the Antarctic margin.
c. Analysis of model output
1) Along-slope velocity component
(a) Map of the topographic gradient, ∇h = (∂h/∂x, ∂h/∂y), calculated from the model’s bathymetry. (b) Ten-year average of along-slope velocity for the upper 500-m water column. Negative velocities circumnavigate Antarctica in a counterclockwise direction. The offshore ocean deeper than 2500 m is masked to set the focus on the circulation near Antarctica.
Citation: Journal of Physical Oceanography 52, 3; 10.1175/JPO-D-21-0143.1
(a) Map of the topographic gradient, ∇h = (∂h/∂x, ∂h/∂y), calculated from the model’s bathymetry. (b) Ten-year average of along-slope velocity for the upper 500-m water column. Negative velocities circumnavigate Antarctica in a counterclockwise direction. The offshore ocean deeper than 2500 m is masked to set the focus on the circulation near Antarctica.
Citation: Journal of Physical Oceanography 52, 3; 10.1175/JPO-D-21-0143.1
(a) Map of the topographic gradient, ∇h = (∂h/∂x, ∂h/∂y), calculated from the model’s bathymetry. (b) Ten-year average of along-slope velocity for the upper 500-m water column. Negative velocities circumnavigate Antarctica in a counterclockwise direction. The offshore ocean deeper than 2500 m is masked to set the focus on the circulation near Antarctica.
Citation: Journal of Physical Oceanography 52, 3; 10.1175/JPO-D-21-0143.1
The along-slope velocity, depth-averaged over the upper 500 m, is shown in Fig. 2b. The ASC emerges as a band of high velocity following the continental slope, as can be seen by comparing the location of the ASC to the location of maximum topographic slope. The ASC is strongest in East Antarctica from the Ross Sea to the Weddell Sea. Even though the topographic gradient at the continental slope is equally large, the ASC is weaker in the Amundsen and Ross Seas and reverses west of the Antarctic Peninsula.
2) Definition of Antarctic Slope Current regimes
We classify the annually averaged along-slope velocities along the 1000-m isobath into three distinct regimes: a surface-intensified ASC, a bottom-intensified ASC, and a reversed ASC. The surface-intensified and bottom-intensified ASC have a negative (counterclockwise around Antarctica) barotropic (depth-averaged), along-slope velocity while the reversed ASC occupies stretches of the margin that have a positive (clockwise around Antarctica) barotropic along-slope velocity, similar to the along-slope velocity averaged over the top 500 m shown in Fig. 2b. The bottom-intensified ASC has velocity values in the bottom 200 m that exceed 50% of the surface 200 m. According to this definition, the surface velocity (upper 200 m) is at least twice the bottom velocity in the surface-intensified ASC.
3) Along- and cross-slope coordinate system
The ASC circumnavigates the Antarctic continent, closely following the continental slope as the large onshore gradient in water column thickness poses a barrier for cross-slope transport (Thompson et al. 2018). The 1000-m isobath provides a continuous contour that approximates the pathway of the ASC (Goddard et al. 2017; Thompson et al. 2018; Moorman et al. 2020; Morrison et al. 2020). However, as the ASC extends across a range of depths, we also transform the ASC velocity data into a new along- and cross-slope coordinate system (see appendix for details) for a more in-depth analysis of the flow field in result sections 3b and 3c. This transformation is designed to unwrap the continental slope by selecting a total of 3520 appropriately chosen cross-slope transects along the entire continental slope (Fig. 3). The model output (e.g., the along-slope velocity) is interpolated onto the transects, which we then arrange next to each other. The approach rectifies the convoluted geometry of the continental slope and the fact that the band occupied by the ASC is very narrow in conventional latitude/longitude coordinates (approximately 50 km in the cross-slope direction) compared to its length (approximately 20 000 km in the along-slope direction).
(a)–(d) Cross-slope transects along the Antarctic continental margin. The black points in (a)–(c) show the location of model grid cells for a smoothed bathymetry along a shelf contour at 630 m and along a deep contour at 2000 m. Gray contours show the nonsmoothed model bathymetry at 500, 1000 (thick), 2000, and 3000 m.
Citation: Journal of Physical Oceanography 52, 3; 10.1175/JPO-D-21-0143.1
(a)–(d) Cross-slope transects along the Antarctic continental margin. The black points in (a)–(c) show the location of model grid cells for a smoothed bathymetry along a shelf contour at 630 m and along a deep contour at 2000 m. Gray contours show the nonsmoothed model bathymetry at 500, 1000 (thick), 2000, and 3000 m.
Citation: Journal of Physical Oceanography 52, 3; 10.1175/JPO-D-21-0143.1
(a)–(d) Cross-slope transects along the Antarctic continental margin. The black points in (a)–(c) show the location of model grid cells for a smoothed bathymetry along a shelf contour at 630 m and along a deep contour at 2000 m. Gray contours show the nonsmoothed model bathymetry at 500, 1000 (thick), 2000, and 3000 m.
Citation: Journal of Physical Oceanography 52, 3; 10.1175/JPO-D-21-0143.1
3. Results
a. Antarctic Slope Current regimes
1) Overview
Conditions influencing the dynamics at the continental shelf break, including dense overflows, surface momentum stress, and the presence of freshwater from basal ice shelf and sea ice melting, vary substantially along the coastline. The spatial variability in forcing leads to a large regional variability in the frontal structure as has been established from both observations (Thompson et al. 2018) and modeling work (Moorman et al. 2020). Assuming geostrophic balance, the shifts in the frontal shape imply that variations also exist in the flow field. We show that the ASC classification indeed largely matches the frontal classification, a link that has previously not been shown with either observational or modeling data in the literature.
Figure 4 shows example transects of annually averaged along-slope velocity, which reveal that the jet structure varies considerably along the continental slope. Figures 4a, 4d, and 4g have a slope jet that flows counterclockwise around Antarctica and is strongest near the surface and subsurface. These transects fall into the first ASC regime, the surface-intensified ASC. The surface-intensified ASC can exist as a narrow jet close to the continental shelf break (Fig. 4a), without a pronounced jet core (Fig. 4d), or as a wide jet with largest velocities on the continental slope (Fig. 4g). In contrast to the surface-intensified ASC of Figs. 4a, 4d, and 4g, the transects of Figs. 4b, 4e, and 4h are examples of the bottom-intensified ASC. Figure 4b shows a clear bottom-intensification with reduced velocities toward the upper water column, Fig. 4e shows additionally a surface-intensified jet that is slightly offset in the offshore direction, and Fig. 4h has the largest velocities of the three that prevail throughout the water column. The third regime we defined shows the characteristics of a reversed ASC (Figs. 4c,f,i). The along-slope velocity changes direction so that the ASC flows in a clockwise direction around Antarctica. The reversed ASC is always surface intensified, and the shelf break constitutes a southern boundary to the flow such that the jet core is located just offshore of the shelf break, with varying intensities between the different transects of Figs. 4c, 4f, and 4i.
(a)–(i) Cross-slope transects of a 10-yr average of along-slope velocity (shading) for (left) the surface-intensified ASC, (center) the bottom-intensified ASC, and (right) the reversed ASC. Black lines are isosurfaces of potential density referenced to the surface (kg m−3). The triangle on top of each plot approximates the position of the 1000-m isobath. (j) Map of the Antarctic coastline. The 1000-m isobath is color-coded accordingly to the ASC regime present on the continental shelf. Numbers near the cross-slope lines correspond to the transects shown in the top panels. Names of regions used in the text are added, where AP is short for Antarctic Peninsula, BS is short for Bellingshausen Sea, and AS is short for Amundsen Sea.
Citation: Journal of Physical Oceanography 52, 3; 10.1175/JPO-D-21-0143.1
(a)–(i) Cross-slope transects of a 10-yr average of along-slope velocity (shading) for (left) the surface-intensified ASC, (center) the bottom-intensified ASC, and (right) the reversed ASC. Black lines are isosurfaces of potential density referenced to the surface (kg m−3). The triangle on top of each plot approximates the position of the 1000-m isobath. (j) Map of the Antarctic coastline. The 1000-m isobath is color-coded accordingly to the ASC regime present on the continental shelf. Numbers near the cross-slope lines correspond to the transects shown in the top panels. Names of regions used in the text are added, where AP is short for Antarctic Peninsula, BS is short for Bellingshausen Sea, and AS is short for Amundsen Sea.
Citation: Journal of Physical Oceanography 52, 3; 10.1175/JPO-D-21-0143.1
(a)–(i) Cross-slope transects of a 10-yr average of along-slope velocity (shading) for (left) the surface-intensified ASC, (center) the bottom-intensified ASC, and (right) the reversed ASC. Black lines are isosurfaces of potential density referenced to the surface (kg m−3). The triangle on top of each plot approximates the position of the 1000-m isobath. (j) Map of the Antarctic coastline. The 1000-m isobath is color-coded accordingly to the ASC regime present on the continental shelf. Numbers near the cross-slope lines correspond to the transects shown in the top panels. Names of regions used in the text are added, where AP is short for Antarctic Peninsula, BS is short for Bellingshausen Sea, and AS is short for Amundsen Sea.
Citation: Journal of Physical Oceanography 52, 3; 10.1175/JPO-D-21-0143.1
The surface-intensified and bottom-intensified ASC regimes occupy a similar amount (9000 and 7500 km, respectively) of the approximately 20 000-km-long 1000-m isobath (Fig. 4j). The reversed ASC spans only 3500 km and is the least common regime.
2) Spatial distribution
Here we compare the ASC regimes, which are a classification based on the velocity field along the 1000-m isobath (Fig. 5), with the previously reported ASF regimes (e.g., Thompson et al. 2018), which are a classification based on cross-slope transects, each 5° apart, of temperature and density [following the method of Moorman et al. (2020)]. The spatial distribution of the ASC regimes aligns well with the switches in the frontal structure, suggesting that the ASC is largely in thermal wind balance (cf. bars in Fig. 5a as well as the velocity field and isopycnals in Fig. 4). The surface-intensified ASC develops in regions where the fresh ASF dominates. The large surface velocities for this regime are in accordance with the large near-surface density gradient existing for fresh shelves from incropping isopycnals. The bottom-intensified ASC occurs where dense shelf waters escape the continental shelf and move downslope. Here, geostrophic adjustment steers the flow to the left, resulting in the along-slope velocity component near the seafloor indicative for this regime. The third ASC regime, the reversed ASC, occurs in the warm shelf regions of the Bellingshausen and Amundsen Seas in West Antarctica where the Antarctic Circumpolar Current extends southward close to the shelf break and influences the ASC. The hydrographic conditions in West Antarctica are in geostrophic balance with the reversal of the flow field (cf. isopycnals and velocity field in Figs. 4c,f,i). An exception to the alignment of the frontal and current regimes is in the Amundsen Sea, between −120° and −150°E, where the onset of the anticlockwise flowing ASC (surface-intensified ASC) is farther east than the transition from warm to fresh shelf. The mismatch in ASF and ASC regimes hints toward a higher sensitivity in this region to regime shifts should the mechanical and/or buoyancy forcing change. Last, the fourth ASF classification, the cold shelf, does not have a corresponding ASC regime and is occupied by either the reversed ASC (tip of the Antarctic Peninsula, west of −55°E) or the surface-intensified ASC (eastern Ross Sea at −170°E).
(a) Total and (b) barotropic along-slope velocity along the 1000-m isobath. For the purpose of this illustration, we filtered the velocity with a running mean of 20 data points. Velocity in (a) is plotted as a function of depth vs distance along the 1000-m isobath and the distance axis is labeled by longitude. Note that this has resulted in stretching of the longitude coordinate in (a) relative to (b) in some locations. The velocity values in (b) are plotted at the location of the 1000-m isobath. Negative velocities are in counterclockwise direction. The bar charts in (a) show a classification for the ASF [the top bar; following the definition of Moorman et al. (2020)] and the ASC (the bottom bar; see text for definition).
Citation: Journal of Physical Oceanography 52, 3; 10.1175/JPO-D-21-0143.1
(a) Total and (b) barotropic along-slope velocity along the 1000-m isobath. For the purpose of this illustration, we filtered the velocity with a running mean of 20 data points. Velocity in (a) is plotted as a function of depth vs distance along the 1000-m isobath and the distance axis is labeled by longitude. Note that this has resulted in stretching of the longitude coordinate in (a) relative to (b) in some locations. The velocity values in (b) are plotted at the location of the 1000-m isobath. Negative velocities are in counterclockwise direction. The bar charts in (a) show a classification for the ASF [the top bar; following the definition of Moorman et al. (2020)] and the ASC (the bottom bar; see text for definition).
Citation: Journal of Physical Oceanography 52, 3; 10.1175/JPO-D-21-0143.1
(a) Total and (b) barotropic along-slope velocity along the 1000-m isobath. For the purpose of this illustration, we filtered the velocity with a running mean of 20 data points. Velocity in (a) is plotted as a function of depth vs distance along the 1000-m isobath and the distance axis is labeled by longitude. Note that this has resulted in stretching of the longitude coordinate in (a) relative to (b) in some locations. The velocity values in (b) are plotted at the location of the 1000-m isobath. Negative velocities are in counterclockwise direction. The bar charts in (a) show a classification for the ASF [the top bar; following the definition of Moorman et al. (2020)] and the ASC (the bottom bar; see text for definition).
Citation: Journal of Physical Oceanography 52, 3; 10.1175/JPO-D-21-0143.1
b. Regime behavior
We have classified the ASC into different regimes based on the vertical structure of the flow field along the 1000-m isobath. The 1000-m isobath is appropriate as the vertical structure of the velocity field is qualitatively the same across the continental slope. However, the depth contour does not necessarily align with the ASC core (cf. triangle location with the velocity field in Figs. 4a–i). To more accurately capture the ASC, we analyze for the remainder of the study characteristics of the three ASC regimes in the along- and cross-slope coordinate system, as introduced in the method section. We determine the ASC regime of each transect in the ASC slope coordinate system using longitude information. Small differences between the two products occur at the tip of the Antarctic Peninsula where undulations of the coastline are not well resolved as the coastline is oriented in a north–south direction, resulting in a comparatively small range of longitude coordinates.
1) Surface, bottom, and barotropic flow
We start by comparing the surface, bottom, and barotropic flow for each regime (Fig. 6). The surface velocity is a depth average over the top 200 m, the bottom velocity is taken from the deepest grid cell, and the barotropic velocity is calculated as the depth average over the water column. The velocity field is spatially highly variable as indicated by the example velocity transects in Fig. 4. By considering regime averages, however, we infer typical characteristics for each regime.
Annual average of (a) upper 200 m (gray), bottom (black), and (b) barotropic along-slope velocity. Velocity data were transformed into an along and cross-slope coordinate system and averaged in the cross-slope direction. Velocity is plotted as a function of transect number and labeled by longitude. Shading in all panels indicates the dominating ASC regime. Over bar values indicate circumpolar regime averages with standard deviations for the upper 200 m (
Citation: Journal of Physical Oceanography 52, 3; 10.1175/JPO-D-21-0143.1
Annual average of (a) upper 200 m (gray), bottom (black), and (b) barotropic along-slope velocity. Velocity data were transformed into an along and cross-slope coordinate system and averaged in the cross-slope direction. Velocity is plotted as a function of transect number and labeled by longitude. Shading in all panels indicates the dominating ASC regime. Over bar values indicate circumpolar regime averages with standard deviations for the upper 200 m (
Citation: Journal of Physical Oceanography 52, 3; 10.1175/JPO-D-21-0143.1
Annual average of (a) upper 200 m (gray), bottom (black), and (b) barotropic along-slope velocity. Velocity data were transformed into an along and cross-slope coordinate system and averaged in the cross-slope direction. Velocity is plotted as a function of transect number and labeled by longitude. Shading in all panels indicates the dominating ASC regime. Over bar values indicate circumpolar regime averages with standard deviations for the upper 200 m (
Citation: Journal of Physical Oceanography 52, 3; 10.1175/JPO-D-21-0143.1
The barotropic velocity is largest for the bottom-intensified ASC with a value of 0.06 m s−1. The large barotropic velocity in the bottom-intensified regime can be explained by the superposition of flows near the bottom and the surface, compared with the surface-intensified ASC where the bottom flow reduces to values close to zero. Note, however, that our definition of the barotropic flow does not imply that the flow in the bottom-intensified ASC (or any other regime) is barotropic in a dynamical sense, meaning that the flow is not necessarily depth-independent.
The increase in the bottom flow is relatively abrupt when approaching a bottom-intensified region from upstream (e.g., at 70°E; Fig. 6a). In contrast, the reduction of the bottom flow downstream of a bottom-intensified region back to values close to zero, characteristic for the surface-intensified ASC, occurs over a longer distance (e.g., west of 55°E). This slow decay occurs because the bottom flow gradually propagates downslope and eventually out of the ASC slope coordinate system. The distance over which the bottom flow decays likely depends on the geometry of the continental slope and the intensity of the overflows upstream. As a result of the slow decay, elevated bottom flows exist in the surface-intensified ASC downstream of a bottom-intensified ASC section (with the exception of the transition from the bottom-intensified ASC to the reversed ASC at the tip of the Antarctic Peninsula at −55°E). These stretches of surface-intensified ASC with larger bottom flows do experience the effect of the dense overflows although they do not fit the criterion for a bottom-intensified ASC. Similarly, bottom-intensified flows also occur away from the ASC, which is farther down the continental slope, in the deep ocean. Examples of a bottom jet, disconnected from the ASC, can be seen in the modeled transect along −246.7°E (113.3°E) of Fig. 1a.
The area of the Antarctic coastline characterized by the reversed ASC has the weakest barotropic flow. The velocities are in particular small when the current changes direction near the transition from the reversed ASC to the surface-intensified ASC at around −120°E. The average surface velocities adjacent to the tip of the Antarctic Peninsula near Drake Passage, however, can be as large as 0.3 m s−1. The transition to the bottom-intensified ASC of the Weddell Sea is abrupt, as conditions profoundly differ on both sides of the Antarctic Peninsula.
The position of the time-averaged current on the continental slope varies between the different regimes. The surface-intensified ASC is closest to the shelf break with maximum barotropic velocities around the 950-m isobath. The bottom-intensified ASC is located farther offshore, around the 1250-m isobath, due to the influence of the dense plume propagation downslope. The reversed ASC is concentrated around the 1100-m isobath.
2) Seasonality
We use the ASC classification based on annual averaged velocities as introduced in section 3a(1) to compare the seasonality of the three different ASC regimes. The approach of having a fixed classification in time for a given location assumes that the regime classification is not sensitive to the seasons. This assumption broadly holds (not shown) with the possible exception of the reversed ASC during winter at the location of the ASC emergence at −130°E, where the ASC changes direction and becomes a westward flowing current, moving eastward along the continental slope. The location of the transition from a bottom-intensified to a surface-intensified ASC downstream of Prydz Bay at around 55°E also varies throughout the year, but is expected to have a minor influence on the overall regime average due to its small area compared with the other bottom-intensified ASC sectors.
The month-to-month variability of the ASC differs between the three regimes (Fig. 7a, showing 3-month rolling averages). Only the surface-intensified ASC has a pronounced seasonal signal in the surface component (solid green line) with velocities 0.1 m s−1 larger in the winter months than in the summer months. Seasonality of the winds may be the relevant driver for the seasonality along stretches of the coastline with a surface-intensified ASC (Fig. 7b). However, the seasonal signal in the ASC is much larger as the along-slope velocities approximately triple, compared with the local surface momentum stress which increases by 50%.
Seasonal cycle of monthly along-slope (a) velocity and (b) surface momentum stress, averaged over each of the three ASC regimes, as well as (c) along-slope velocity for the bottom-intensified ASC sector near Prydz Bay. All data were transformed into an along- and cross-slope coordinate system and averaged in the cross-slope direction. All seasonal cycles are filtered with a 3-month running mean.
Citation: Journal of Physical Oceanography 52, 3; 10.1175/JPO-D-21-0143.1
Seasonal cycle of monthly along-slope (a) velocity and (b) surface momentum stress, averaged over each of the three ASC regimes, as well as (c) along-slope velocity for the bottom-intensified ASC sector near Prydz Bay. All data were transformed into an along- and cross-slope coordinate system and averaged in the cross-slope direction. All seasonal cycles are filtered with a 3-month running mean.
Citation: Journal of Physical Oceanography 52, 3; 10.1175/JPO-D-21-0143.1
Seasonal cycle of monthly along-slope (a) velocity and (b) surface momentum stress, averaged over each of the three ASC regimes, as well as (c) along-slope velocity for the bottom-intensified ASC sector near Prydz Bay. All data were transformed into an along- and cross-slope coordinate system and averaged in the cross-slope direction. All seasonal cycles are filtered with a 3-month running mean.
Citation: Journal of Physical Oceanography 52, 3; 10.1175/JPO-D-21-0143.1
To further complicate the relationship between the ASC seasonality and the winds, the bottom-intensified and reversed ASC regimes also experience larger surface momentum stress in the winter months; however, the ASC in the reversed and bottom-intensified regimes does not respond in a similar way. The bottom-intensified ASC experiences a similar increase in the surface momentum stress (70% increase) in the winter months as the surface-intensified ASC, and the seasonal signal of the surface velocity varies with a 60% increase by an equal amount. The reversed ASC has an inverted seasonality, meaning it is weaker in the winter months. A likely explanation is the strengthening of the along-slope surface momentum stress component (Fig. 7b), which opposes the current. Thus, the stronger easterly winds in winter in this region will slow the current down, or in some regions act to reverse the flow toward an anticlockwise flowing current (not shown). Any variability of the bottom-intensified ASC sections can be expected to be modulated by the timing of overflows. The weak seasonality in the modeled overflows (Morrison et al. 2020; Fig. 4) results in the weak seasonal signal of Fig. 7a. An exception is the bottom-intensified ASC section downstream of Prydz Bay (Fig. 7c). Here, the surface and bottom flow both vary, but with a different timing. The surface velocity peaks in May, similar to the surface flow of the neighboring surface-intensified ASC sections. The bottom flow is strongest in winter co-occurring with the surface intensification, which suggests a barotropic response. A second peak exists in November that is likely related to a higher frequency of overflows occurring at the end of winter and the dense shelf water production season (not shown).
c. Antarctic Slope Current perturbations
To investigate higher temporal variability of the ASC, we now consider daily along-slope velocities averaged in the cross-slope direction in a Hovmöller diagram (Fig. 8). Perturbations, such as eddies, that propagate in time along the continental slope emerge in the surface velocity (Fig. 8a) as well as in the barotropic velocity (Fig. 8b), as can be seen by the angled bands that track the propagation of features along the slope. Our goal is to relate the along-slope propagation of the ASC perturbations to individual processes. Note, however, that other features of the ASC that we presented earlier in this study can also be seen in Fig. 8: for example, the high spatial variability of the along-slope velocities [discussed in section 3b(1)] or the seasonality of the ASC with a reversal of the flow in the Amundsen Sea, between −120°E and −150°E [discussed in section 3a(2) on the spatial distribution], and the higher velocities in the winter months in the surface-intensified ASC while little seasonality exists in the bottom-intensified or reversed ASC [discussed in section 3b(2)].
Hovmöller diagram of (a) surface and (b) barotropic daily along-slope velocity along the Antarctic continental slope showing the propagation of velocity perturbations. Velocity data were transformed into an along- and cross-slope coordinate system and averaged in cross-slope direction. Velocity is plotted as a function of transect number and labeled by longitude. The yellow lines indicate the characteristic propagation speed of wind perturbations as well as of the barotropic and bottom velocity, respectively. The two black lines have the same slope and are added to highlight the existence of additional time scales on which signals travel along the continental slope. The steeper the slope of the lines, the slower the signal propagation (propagation speed is shown next to each line). The bar chart on top of the figure shows the ASC regime classification (see text for definition) in the along- and cross-slope coordinate system.
Citation: Journal of Physical Oceanography 52, 3; 10.1175/JPO-D-21-0143.1
Hovmöller diagram of (a) surface and (b) barotropic daily along-slope velocity along the Antarctic continental slope showing the propagation of velocity perturbations. Velocity data were transformed into an along- and cross-slope coordinate system and averaged in cross-slope direction. Velocity is plotted as a function of transect number and labeled by longitude. The yellow lines indicate the characteristic propagation speed of wind perturbations as well as of the barotropic and bottom velocity, respectively. The two black lines have the same slope and are added to highlight the existence of additional time scales on which signals travel along the continental slope. The steeper the slope of the lines, the slower the signal propagation (propagation speed is shown next to each line). The bar chart on top of the figure shows the ASC regime classification (see text for definition) in the along- and cross-slope coordinate system.
Citation: Journal of Physical Oceanography 52, 3; 10.1175/JPO-D-21-0143.1
Hovmöller diagram of (a) surface and (b) barotropic daily along-slope velocity along the Antarctic continental slope showing the propagation of velocity perturbations. Velocity data were transformed into an along- and cross-slope coordinate system and averaged in cross-slope direction. Velocity is plotted as a function of transect number and labeled by longitude. The yellow lines indicate the characteristic propagation speed of wind perturbations as well as of the barotropic and bottom velocity, respectively. The two black lines have the same slope and are added to highlight the existence of additional time scales on which signals travel along the continental slope. The steeper the slope of the lines, the slower the signal propagation (propagation speed is shown next to each line). The bar chart on top of the figure shows the ASC regime classification (see text for definition) in the along- and cross-slope coordinate system.
Citation: Journal of Physical Oceanography 52, 3; 10.1175/JPO-D-21-0143.1
The most prominent feature in the surface velocity field is a very fast signal propagating clockwise around Antarctica (Fig. 8a). We can relate the propagation to the time scale of perturbations in the easterly winds, which take approximately 23 days to circumnavigate the continent in the JRA55-do v1.3 atmospheric forcing product. The yellow lines, added at two different locations, represent the propagation speed of the wind perturbations. The slope of the yellow lines matches well the slope of the fast propagating high and low velocity signal in the surface velocity field. The wind perturbations are a signature of the high and low pressure systems (storms) from the westerlies at lower latitudes and travel in eastward direction, even though the expression of the winds at the ocean surface is westward.
The propagation speed of many of the perturbations away from the surface can be related to the advection time scale of the current itself (highlighted with yellow lines in Fig. 8b). Consequently, the angle of the lines with the horizontal varies along the continental slope and the direction differs between the bottom-intensified and surface-intensified ASC (both propagating counterclockwise around Antarctica) and the reversed ASC (propagating clockwise around Antarctica). For the bottom-intensified ASC, it is the advection time scale of the bottom velocity, instead of the barotropic velocity, that matches best the slope of the perturbations in the Hovmöller diagram. The perturbations in this regime, most pronounced in the Ross Sea at around −180°, likely originate from the dense overflows that set the frequency for the perturbations, which is higher here than in the neighboring regions (Gordon et al. 2009).
Note that advection is not the only process that determines how perturbations travel along the continental slope. Various additional angled bands exist, as shown by way of example by the two black lines in Fig. 8b. The black lines correspond to a signal that travels with approximately 41 km day−1 westward along the continental slope (i.e., the shallow side is on its left) and that might be connected to topographic Rossby waves (Zhao and Timmermans 2018).
4. Discussion
We have provided a comprehensive description of the along-slope velocity structure along the Antarctic continental slope in an eddying ocean–sea ice model and revealed the large-scale spatial variability by categorizing the ASC into three regimes: the surface-intensified ASC, the bottom-intensified ASC, and the reversed ASC (Fig. 9). Each ASC regime corresponds to a distinct classification of the density field as introduced by Thompson et al. (2018). The surface-intensified ASC occurs at sections of the continental slope with a fresh shelf, the bottom-intensified ASC occurs at dense shelves, and the reversed ASC occurs at warm shelves.
Schematic representation of three distinct regimes of the ASC system at the Antarctic continental shelf break. Black lines approximate the characteristic shape of the isopycnals, which (a) incrops with the bathymetry for the fresh ASF, (b) follows a “V” shape for the dense ASF, and (c) is approximately horizontal for the warm ASF. The shading indicates the ASC where red shading indicates flow out of the page (reversed ASC) and blue shading indicates flow into the page (bottom-intensified and surface-intensified ASC). Land is shown in brown and the ocean surface by the blue line.
Citation: Journal of Physical Oceanography 52, 3; 10.1175/JPO-D-21-0143.1
Schematic representation of three distinct regimes of the ASC system at the Antarctic continental shelf break. Black lines approximate the characteristic shape of the isopycnals, which (a) incrops with the bathymetry for the fresh ASF, (b) follows a “V” shape for the dense ASF, and (c) is approximately horizontal for the warm ASF. The shading indicates the ASC where red shading indicates flow out of the page (reversed ASC) and blue shading indicates flow into the page (bottom-intensified and surface-intensified ASC). Land is shown in brown and the ocean surface by the blue line.
Citation: Journal of Physical Oceanography 52, 3; 10.1175/JPO-D-21-0143.1
Schematic representation of three distinct regimes of the ASC system at the Antarctic continental shelf break. Black lines approximate the characteristic shape of the isopycnals, which (a) incrops with the bathymetry for the fresh ASF, (b) follows a “V” shape for the dense ASF, and (c) is approximately horizontal for the warm ASF. The shading indicates the ASC where red shading indicates flow out of the page (reversed ASC) and blue shading indicates flow into the page (bottom-intensified and surface-intensified ASC). Land is shown in brown and the ocean surface by the blue line.
Citation: Journal of Physical Oceanography 52, 3; 10.1175/JPO-D-21-0143.1
(a)–(c) Bathymetry profiles for three example transects where the bathymetry is nonmonotonic. The continental slope is represented by a step function because the model bathymetry is interpolated onto 50 regularly spaced points in the cross-slope direction, while the continental slope is represented by a varying number (<50) of grid points in the model. The two horizontal black lines indicate the 650- and 2000-m depths. In (a) the bathymetry profile has undulations smaller than 30 m (blue), which are smoothed (black); all other points in the profile are kept (red). In (b) the bathymetry profile has undulations between 30 and 200 m (blue), which are flattened out (black); all other points in the profile are kept (red). In (c) the bathymetry profile has undulations larger than 200 m. These profiles are deleted; for their location see (d). (d) Cross-slope transects along the Antarctic continental margin. Transects highlighted in blue are removed as they have large undulations in their bathymetry profile and are not monotonic [example in (c)]. Gray contours show the nonsmoothed model bathymetry at 500, 1000 (thick), 2000, and 3000 m.
Citation: Journal of Physical Oceanography 52, 3; 10.1175/JPO-D-21-0143.1
(a)–(c) Bathymetry profiles for three example transects where the bathymetry is nonmonotonic. The continental slope is represented by a step function because the model bathymetry is interpolated onto 50 regularly spaced points in the cross-slope direction, while the continental slope is represented by a varying number (<50) of grid points in the model. The two horizontal black lines indicate the 650- and 2000-m depths. In (a) the bathymetry profile has undulations smaller than 30 m (blue), which are smoothed (black); all other points in the profile are kept (red). In (b) the bathymetry profile has undulations between 30 and 200 m (blue), which are flattened out (black); all other points in the profile are kept (red). In (c) the bathymetry profile has undulations larger than 200 m. These profiles are deleted; for their location see (d). (d) Cross-slope transects along the Antarctic continental margin. Transects highlighted in blue are removed as they have large undulations in their bathymetry profile and are not monotonic [example in (c)]. Gray contours show the nonsmoothed model bathymetry at 500, 1000 (thick), 2000, and 3000 m.
Citation: Journal of Physical Oceanography 52, 3; 10.1175/JPO-D-21-0143.1
(a)–(c) Bathymetry profiles for three example transects where the bathymetry is nonmonotonic. The continental slope is represented by a step function because the model bathymetry is interpolated onto 50 regularly spaced points in the cross-slope direction, while the continental slope is represented by a varying number (<50) of grid points in the model. The two horizontal black lines indicate the 650- and 2000-m depths. In (a) the bathymetry profile has undulations smaller than 30 m (blue), which are smoothed (black); all other points in the profile are kept (red). In (b) the bathymetry profile has undulations between 30 and 200 m (blue), which are flattened out (black); all other points in the profile are kept (red). In (c) the bathymetry profile has undulations larger than 200 m. These profiles are deleted; for their location see (d). (d) Cross-slope transects along the Antarctic continental margin. Transects highlighted in blue are removed as they have large undulations in their bathymetry profile and are not monotonic [example in (c)]. Gray contours show the nonsmoothed model bathymetry at 500, 1000 (thick), 2000, and 3000 m.
Citation: Journal of Physical Oceanography 52, 3; 10.1175/JPO-D-21-0143.1
Direct observations of the vertical velocity structure along the continental slope are sparse, but available snapshots in time or point measurements of the velocity field qualitatively match the modeled large-scale variability of the ASC. For example, surface-intensified velocity profiles have also been observed in the eastern Weddell Sea (Heywood et al. 1998; Chavanne et al. 2010) and bottom-intensified velocities in the western Weddell Sea (Gordon et al. 2009; Azaneu et al. 2017) and in the Prydz Bay region west of Cape Darnley (Fukamachi et al. 2000; Hirano et al. 2015). The mooring transect of PM2016, which we used to qualitatively evaluate the model, is near the transition zone we identified between the bottom-intensified ASC of Adélie Land and the surface-intensified ASC to the west and shows a bottom-intensification in the observations. The modeled transect shows surface-intensified velocities with a bottom-intensified flow farther down the slope (Fig. 1a), highlighting the sensitivity of ASC regimes to the exact pathways of dense bottom waters. The initiation of the counterclockwise flowing and surface-intensified ASC in West Antarctica develops in the model at −120°E, farther west than where observations suggest (Walker et al. 2013; Thompson et al. 2020). The offset of the ASC emergence in the model to the west is consistent with a warm bias of the continental shelf temperatures and a mismatch of the modeled warm shelf compared to an observed fresh shelf (Moorman et al. 2020; Thompson et al. 2018). The comparison of the model to observations, however, comes with the caveat of limited profiles taken in this region.
The classification into regimes allows us to expose differences in the seasonality of the ASC along the continental slope. In particular, we showed that only the surface-intensified ASC has a notable seasonal signal. Winds have previously been suggested to drive the seasonality of the ASC (Núñez-Riboni and Fahrbach 2009; Peña-Molino et al. 2016; Armitage et al. 2018) and while we cannot exclude winds as an important driver, winds alone do not explain why only the surface-intensified ASC varies substantially throughout the year as the bottom-intensified and reversed ASC regimes are equally exposed to a seasonally changing surface momentum stress (Fig. 7b). While we are unable to further test the effect of winds in this study, perturbation experiments in which the strength of the easterly winds is changed would be a way to assess the role of winds with regard to the ASC seasonality in the future.
Seasonality of the ASC is influenced not only by the mechanical forcing provided by the winds (and sea ice) at the ocean surface, but also by a geostrophic adjustment to changes in the cross-slope density gradient via freshwater input from basal melting and via surface water mass transformation. Seasonal variability in the cross-slope density gradient computed at 300-m depth from observations in Pauthenet et al. (2021) show largest gradients in summer when continental shelf salinities are minimal, suggesting a winter slowdown of the baroclinic component of the near-surface flow. The prescribed freshwater input from basal melting in the model is constant in time, which introduces an error (Hattermann et al. 2012; Herraiz-Borreguero et al. 2015) possibly resulting in a misrepresentation of the ASC seasonality in some areas. In contrast, the model is able to simulate realistic surface water transformation rates. Bottom water is exported down the continental slope throughout the year in the western Weddell Sea, the Ross Sea, and Adélie Land, and has a seasonal cycle in Prydz Bay, similar to what has been reported from observations (Gill 1973; Gordon et al. 2009; Williams et al. 2010; Fukamachi et al. 2000).
The model’s ability to produce and export bottom water is crucial to reproduce the bottom-intensified ASC regime. However, the model lacks some physics, such as the aforementioned thermodynamic interaction with ice shelves, tides, and full mesoscale and submesoscale dynamics, due to insufficient grid resolution. The grid resolution, in particular, may explain why the ASC forms in the model as a relatively broad jet instead of multiple narrow jets that drift across the continental slope, as high-resolution regional modeling studies as well as observational studies reported (Stern et al. 2015; Peña-Molino et al. 2016; Azaneu et al. 2017).
Last, reversals in the direction of the bottom flow compared to the direction of the surface flow occur frequently on individual days due to an unknown mechanism throughout the year for the surface-intensified and the reversed ASC (123 and 101 days, respectively) for which the bottom velocities reduce to average values of 0.1 and −0.1 m s−1, respectively. The finding of a fast changing ASC could be related to topographic Rossby waves. Changes in the ASC on short time scales are also supported by a velocity transect from the eastern Weddell Sea that shows an undercurrent the first time the transect was observed, but lacks the undercurrent when the transect was repeated a month later (Chavanne et al. 2010). In contrast, Silvano et al. (2019) propose that the undercurrent at approximately −242°E (118°E) in East Antarctica is a summer feature, rather than a high-frequency event occurring throughout the year.
5. Conclusions
We have provided a comprehensive description of the ASC’s spatial and subannual variability in an ocean–sea ice model. We divided the ASC into three distinct regimes that align spatially well with frontal regime classifications, suggesting that the velocity and density field are governed by the same leading-order dynamics around the Antarctic continental slope (Figs. 5 and 9). By classifying the velocity field into different regimes we reveal the large-scale variability of the ASC. Another important result is that the ASC additionally varies within each regime on much smaller spatial scales (Figs. 4 and 6). Our results further showed that only the surface-intensified ASC has a pronounced seasonality (Fig. 7). Subannual variability is, however, not limited to the seasonal time scale, as perturbations of the ASC travel with varying time scales along the continental slope (Fig. 8).
These results have raised a number of questions on the dynamics of the ASC for future studies to address: What controls the seasonal variability in the different ASC regimes and how does the ASC vary on interannual time scales? What is the role of the continental slope steepness in setting the small-scale spatial variability of the ASC? How is the reversed ASC maintained when the surface momentum stress is in the opposite direction? How coupled are the surface and bottom flows in the bottom-intensified ASC and what will be the impact of a possible reduction in dense overflows in a future climate? We emphasize that closing these knowledge gaps is important in our ability to understand the region’s response to a changing climate.
Acknowledgments.
This research was undertaken on the National Computational Infrastructure (NCI) in Canberra, Australia, which is supported by the Australian Commonwealth Government. All authors were supported by the ARC Discovery Project DP190100494. AKM was supported by an Australian Research Council DECRA Fellowship DE170100184. We thank the COSIMA consortium (http://cosima.org.au) for technical support and the development of the ACCESS-OM2-01 model configuration used.
Data availability statement.
The model output on which this paper is based is too large to be retained or publicly archived with available resources. Post-processed model output is available at the Zenodo repository (https://doi.org/10.5281/zenodo.5077402). The observational datasets for this research are included in Peña-Molino et al. (2016). Analysis and figures included in this paper can be reproduced using Jupyter notebooks that are available at the GitHub repository https://github.com/wghuneke/ASC_SpatialTemporalVariability.
APPENDIX
Along- and Cross-Slope Coordinate System
For the coordinate transformation, we interpolate the model output onto cross-slope transects that cover the entire continental slope. The data product is used in Figs. 6, 7, and 8. The cross-slope transects are obtained as follows. We start by selecting a shelf isobath at 630 m and a deep isobath at 2000 m. The shelf and deep isobaths are chosen such that they bound the ASC along most of the continental slope, while being isobaths that are not too convoluted. A shallower isobath than 630 m would involve extensions all the way to the coastline in locations where troughs connect the shelf break with the inner continental shelf. Similarly, isobaths deeper than 2000 m extend in some sectors much farther into the Southern Ocean and away from the ASC. We select both contours, those for the shelf and the deep isobath, for a bathymetry that we first smooth horizontally over 10 grid cells in both directions. The rationale for the smoothing is to avoid complications from complex fine-scale bathymetric features when finding the cross-slope orientation of the transects. We use the smoothed bathymetry only for the selection of the cross-slope transects. All further computations, such as the rotation of the velocity field into along- and cross-slope components and the interpolation of the velocity field onto the transects explained further below, use the nonsmoothed model grid data.
To obtain the cross-slope transects, we start at a point on the shelf isobath and find the point on the deep isobath with the shortest distance from the shelf isobath. We repeat the shortest distance selection for every tenth grid point on the shelf isobath. Additional transects are obtained by evenly distributing transects between the main transects found in the first step (Fig. 3). Directly choosing all transects per shortest distance selection, starting from the shelf isobath, results in a low spatial coverage as multiple transects end at the same point on the deep isobath, which creates large gaps between the transects at the deep end. The present approach therefore represents a compromise between a true cross-slope orientation of the transects and a good spatial coverage of the continental slope. Slight deviations of the transects from the true cross-slope direction do not impact the results because the velocities that we interpolate onto these transects have been rotated into the true along-slope direction, using the unsmoothed bathymetry. We hand-edit or remove any transects that overlap or are not approximating the cross-slope direction, which sometimes occurs near narrow troughs or spurs.
The along-slope velocity component on the nonsmoothed bathymetry grid is then interpolated horizontally onto the transects using the nearest value method and 50 regularly spaced points in the cross-slope direction. In a final step, each transect is interpolated horizontally along the transect onto discrete isobath coordinate values. The isobath coordinates range between 650 and 2000 m and vary in size along the transect as they are set to be double the vertical model grid, meaning there are two bins for each vertical grid cell. For the interpolation onto the isobath coordinate, we use a linear interpolation for the ocean interior and the nearest values method near the bottom to ensure no bottom data are lost in the interpolation step.
The interpolation onto the isobath coordinate requires the bathymetry of the cross-slope transects to be monotonic. For a number of transects the bathymetry does not steadily deepen; this is an artifact of the cross-slope transect selection on a smoothed bathymetry and due to the continental slope’s many undulations. We therefore examine each transect for monotonicity and manipulate the bathymetry, if required, before interpolating onto the isobath coordinate, as follows. We delete transects where the shallowest value is deeper than 1200 m (126 transects), which occurs where the continental shelf is narrower than a single model grid cell and the bathymetry deepens abruptly. Three cases emerge for the remaining, nonmonotonic transects (Figs. A1a–c, where red indicates points that are kept, blue indicates points that are nonmonotonic, and black indicates the updated depth value of the blue points): (i) transects with undulations in the bathymetry that are smaller than 30 m are smoothed (322 transects), (ii) transects with undulations between 30 and 200 m have the undulations either flattened (395 transects) or truncated if the undulation coincides with the shallowest or deepest point of the transect (147 transects), and (iii) transects with undulations larger than 200 m are deleted (121 transects). All deleted transects (7% of the total number of transects) are marked in blue in Fig. A1d, showing they predominantly occur near troughs and spurs as well as around the prime meridian where the continental shelf is very narrow.
REFERENCES
Armitage, T. W., R. Kwok, A. F. Thompson, and G. Cunningham, 2018: Dynamic topography and sea level anomalies of the Southern Ocean: Variability and teleconnections. J. Geophys. Res. Oceans, 123, 613–630, https://doi.org/10.1002/2017JC013534.
Azaneu, M. V. C., K. J. Heywood, B. Y. Queste, and A. F. Thompson, 2017: Variability of the Antarctic Slope Current system in the northwestern Weddell Sea. J. Phys. Oceanogr., 47, 2977–2997, https://doi.org/10.1175/JPO-D-17-0030.1.
Baines, P. G., 2009: A model for the structure of the Antarctic Slope Front. Deep-Sea Res. II, 56, 859–873, https://doi.org/10.1016/j.dsr2.2008.10.030.
Chavanne, C. P., K. J. Heywood, K. W. Nicholls, and I. Fer, 2010: Observations of the Antarctic Slope Undercurrent in the southeastern Weddell Sea. Geophys. Res. Lett., 37, L13601, https://doi.org/10.1029/2010GL043603.
DeConto, R. M., and D. Pollard, 2016: Contribution of Antarctica to past and future sea-level rise. Nature, 531, 591–597, https://doi.org/10.1038/nature17145.
Depoorter, M. A., J. L. Bamber, J. A. Griggs, J. T. M. Lenaerts, S. R. M. Ligtenberg, M. R. van den Broeke, and G. Moholdt, 2013: Calving fluxes and basal melt rates of Antarctic ice shelves. Nature, 502, 89–92, https://doi.org/10.1038/nature12567.
Dinniman, M. S., J. M. Klinck, and E. E. Hofmann, 2012: Sensitivity of Circumpolar Deep Water transport and ice shelf basal melt along the West Antarctic Peninsula to changes in the winds. J. Climate, 25, 4799–4816, https://doi.org/10.1175/JCLI-D-11-00307.1.
Flexas, M. D. M., M. P. Schodlok, L. Padman, D. Menemenlis, and A. H. Orsi, 2015: Role of tides on the formation of the Antarctic Slope Front at the Weddell-Scotia Confluence. J. Geophys. Res. Oceans, 120, 3658–3680, https://doi.org/10.1002/2014JC010372.
Fukamachi, Y., M. Wakatsuchi, K. Taira, S. Kitagawa, T. Furukawa, and M. Fukuchi, 2000: Seasonal variability of bottom water properties off Adélie Land, Antarctica. J. Climate, 105, 6531–6540, https://doi.org/10.1029/1999JC900292.
GEBCO, 2014: GEBCO2014 grid. GEBCO, https://www.gebco.net/news_and_media/gebco_2014_grid.html.
Gill, A. E., 1973: Circulation and bottom water production in the Weddell Sea. Deep-Sea Res., 20, 111–140, https://doi.org/10.1016/0011-7471(73)90048-X.
Goddard, P. B., C. O. Dufour, J. Yin, S. M. Griffies, and M. Winton, 2017: CO2-induced ocean warming of the Antarctic continental shelf in an eddying global climate model. J. Geophys. Res. Oceans, 122, 8079–8101, https://doi.org/10.1002/2017JC012849.
Gordon, A. L., A. H. Orsi, R. Muench, B. A. Huber, E. Zambianchi, and M. Visbeck, 2009: Western Ross Sea continental slope gravity currents. Deep-Sea Res. II, 56, 796–817, https://doi.org/10.1016/j.dsr2.2008.10.037.
Griffies, S. M., 2012: Elements of the Modular Ocean Model (MOM): 2012 release. GFDL Ocean Group Tech. Rep. 7, NOAA Geophysical Fluid Dynamics Laboratory, 618 pp.
Hattermann, T., O. A. Nost, J. M. Lilly, and L. H. Smedsrud, 2012: Two years of oceanic observations below the Fimbul Ice Shelf, Antarctica. Geophys. Res. Lett., 39, L12605, https://doi.org/10.1029/2012GL051012.
Hattermann, T., L. H. Smedsrud, O. A. Nøst, J. M. Lilly, and B. K. Galton-Fenzi, 2014: Eddy-resolving simulations of the Fimbul Ice Shelf cavity circulation: Basal melting and exchange with open ocean. Ocean Modell., 82, 28–44, https://doi.org/10.1016/j.ocemod.2014.07.004.
Herraiz-Borreguero, L., R. Coleman, I. Allison, S. R. Rintoul, M. Craven, and G. D. Williams, 2015: Circulation of modified Circumpolar Deep Water and basal melt beneath the Amery Ice Shelf, East Antarctica. J. Geophys. Res. Oceans, 120, 3098–3112, https://doi.org/10.1002/2015JC010697.
Heywood, K. J., R. A. Locarnini, R. D. Frew, P. F. Dennis, and B. A. King, 1998: Transport and water masses of the Antarctic Slope Front system in the eastern Weddell Sea. Ocean, Ice and Atmosphere: Interactions at the Antarctic Continental Margin, S. S. Jacobs and R. F. Weiss, Eds., Antarctic Research Series, Vol. 75, Wiley, 203–214.
Hirano, D., Y. Kitade, K. I. Ohshima, and Y. Fukamachi, 2015: The role of turbulent mixing in the modified Shelf Water overflows that produce Cape Darnley Bottom Water. J. Geophys. Res. Oceans, 120, 910–922, https://doi.org/10.1002/2014JC010059.
Hunke, E. C., W. H. Lipscomb, A. K. Turner, N. Jeffery, and S. Elliott, 2015: CICE: The Los Alamos sea ice model documentation and software user’s manual. Tech. Rep., Los Alamos National Laboratory, 116 pp.
Jacobs, S. S., 1991: On the nature and significance of the Antarctic Slope Front. Mar. Chem., 35, 9–24, https://doi.org/10.1016/S0304-4203(09)90005-6.
Jacobs, S. S., 2004: Bottom water production and its links with the thermohaline circulation. Antarct. Sci., 16, 427–437, https://doi.org/10.1017/S095410200400224X.
Kiss, A. E., and Coauthors, 2020: ACCESS-OM2: A global ocean–sea ice model at three resolutions. Geosci. Model Dev. Discuss., 13, 401–442, https://doi.org/10.5194/gmd-13-401-2020.
Mathiot, P., H. Goosse, T. Fichefet, B. Barnier, and H. Gallée, 2011: Modelling the seasonal variability of the Antarctic Slope Current. Ocean Sci., 7, 455–470, https://doi.org/10.5194/os-7-455-2011.
Moorman, R., A. K. Morrison, and A. M. Hogg, 2020: Thermal responses to Antarctic ice shelf melt in an eddy-rich global ocean–sea-ice model. J. Climate, 33, 6599–6620, https://doi.org/10.1175/JCLI-D-19-0846.1.
Morrison, A. K., A. M. Hogg, M. H. England, and P. Spence, 2020: Warm circumpolar deep water transport toward Antarctica driven by local dense water export in canyons. Sci. Adv., 6, eaav2516, https://doi.org/10.1126/sciadv.aav2516.
Nøst, O. A., M. Biuw, V. Tverberg, C. Lydersen, T. Hattermann, Q. Zhou, L. H. Smedsrud, and K. M. Kovacs, 2011: Eddy overturning of the Antarctic Slope Front controls glacial melting in the eastern Weddell Sea. J. Geophys. Res. Oceans, 116, C11014, https://doi.org/10.1029/2011JC006965.
Núñez-Riboni, I., and E. Fahrbach, 2009: Seasonal variability of the Antarctic Coastal Current and its driving mechanisms in the Weddell Sea. Deep-Sea Res. I, 56, 1927–1941, https://doi.org/10.1016/j.dsr.2009.06.005.
Pauthenet, E., J.-B. Sallée, S. Schmidtko, and D. Nerini, 2021: Seasonal variation of the Antarctic Slope Front occurrence and position estimated from an interpolated hydrographic climatology. J. Phys. Oceanogr., 51, 1539–1557, https://doi.org/10.1175/JPO-D-20-0186.1.
Peña-Molino, B., M. S. McCartney, and S. R. Rintoul, 2016: Direct observations of the Antarctic Slope Current transport at 113°E. J. Geophys. Res. Oceans, 121, 7390–7407, https://doi.org/10.1002/2015JC011594.
Silvano, A., S. R. Rintoul, K. Kusahara, B. Peña‐Molino, E. Wijk, D. E. Gwyther, and G. D. Williams, 2019: Seasonality of warm water intrusions onto the continental shelf near the Totten Glacier. J. Geophys. Res. Oceans, 124, 4272–4289, https://doi.org/10.1029/2018JC014634.
Stern, A., L.-P. Nadeau, and D. Holland, 2015: Instability and mixing of zonal jets along an idealized continental shelf break. J. Phys. Oceanogr., 45, 2315–2338, https://doi.org/10.1175/JPO-D-14-0213.1.
Stewart, A. L., and A. F. Thompson, 2015: Eddy-mediated transport of warm Circumpolar Deep Water across the Antarctic shelf break. Geophys. Res. Lett., 42, 432–440, https://doi.org/10.1002/2014GL062281.
Stewart, A. L., A. Klocker, and D. Menemenlis, 2019: Acceleration and overturning of the Antarctic Slope Current by winds, eddies, and tides. J. Phys. Oceanogr., 49, 2043–2074, https://doi.org/10.1175/JPO-D-18-0221.1.
Stewart, K. D., and Coauthors, 2020: JRA55-do-based repeat year forcing datasets for driving ocean–sea-ice models. Ocean Modell., 147, 101557, https://doi.org/10.1016/j.ocemod.2019.101557.
St-Laurent, P., J. M. Klinck, and M. S. Dinniman, 2013: On the role of coastal troughs in the circulation of warm Circumpolar Deep Water on Antarctic shelves. J. Phys. Oceanogr., 43, 51–64, https://doi.org/10.1175/JPO-D-11-0237.1.
Thompson, A. F., A. L. Stewart, P. Spence, and K. J. Heywood, 2018: The Antarctic Slope Current in a changing climate. Rev. Geophys., 56, 741–770, https://doi.org/10.1029/2018RG000624.
Thompson, A. F., K. G. Speer, and L. M. S. Chretien, 2020: Genesis of the Antarctic Slope Current in West Antarctica. Geophys. Res. Lett., 47, e2020GL087802, https://doi.org/10.1029/2020GL087802.
Tsujino, H., and Coauthors, 2018: JRA-55 based surface dataset for driving ocean–sea-ice models (JRA55-do). Ocean Modell., 130, 79–139, https://doi.org/10.1016/j.ocemod.2018.07.002.
Walker, D. P., A. Jenkins, K. M. Assmann, D. R. Shoosmith, and M. A. Brandon, 2013: Oceanographic observations at the shelf break of the Amundsen Sea, Antarctica. J. Geophys. Res. Oceans, 118, 2906–2918, https://doi.org/10.1002/jgrc.20212.
Whitworth, T., A. H. Orsi, S.-J. Kim, W. D. Nowlin, and R. A. Locarnini, 1998: Water masses and mixing near the Antarctic slope front. Ocean, Ice and Atmosphere: Interactions at the Antarctic Continental Margin, S. S. Jacobs and R. F. Weiss, Eds., Antarctic Research Series, Vol. 75, Wiley, 1–27, https://doi.org/10.1029/AR075p0001.
Williams, G. D., S. Aoki, S. S. Jacobs, S. R. Rintoul, T. Tamura, and N. L. Bindoff, 2010: Antarctic Bottom Water from the Adélie and George V Land coast, East Antarctica (140–149°E). J. Geophys. Res. Oceans, 115, C04027, https://doi.org/10.1029/2009JC005812.
Zhao, B., and M. L. Timmermans, 2018: Topographic Rossby waves in the Arctic Ocean’s Beaufort Gyre. J. Geophys. Res. Oceans, 123, 6521–6530, https://doi.org/10.1029/2018JC014233.
