• Foster, T. D., and E. C. Carmack, 1976: Temperature and salinity structure in the Weddell Sea. J. Phys. Oceanogr., 6, 3644.

  • Gill, A. E., 1973: Circulation and bottom water production in the Weddell Sea. Deep-Sea Res., 20, 111140.

  • Gordon, A. L., 1971: Oceanography of Antarctic waters. Antarctic Oceanology, Vol. 1, J. L. Reid, Ed., Antarctic Research Series, Vol. 15, Amer. Geophys. Union, 169–203.

    • Search Google Scholar
    • Export Citation
  • Gordon, A. L., 1981: Seasonality of Southern Ocean sea ice. J. Geophys. Res., 86 (C5), 41934197.

  • Gordon, A. L., and B. A. Huber, 1990: Southern Ocean winter mixed layer. J. Geophys. Res., 95 (C7), 11 65511 672.

  • Gordon, A. L., C. T. A. Chen, and W. G. Metcalf, 1984: Winter mixed layer entrainment of Weddell Deep Water. J. Geophys. Res., 89 (C1), 637640.

    • Search Google Scholar
    • Export Citation
  • Huppert, H. E., 1971: On the stability of a series of double-diffusive layers. Deep-Sea Res., 18, 10051021.

  • Klatt, O., O. Boebel, and E. Fahrbach, 2007: A profiling float’s sense of ice. J. Atmos. Oceanic Technol., 24, 13011308.

  • Lytle, V. I., R. Massom, N. Bindoff, A. Worby, and I. Allison, 2000: Wintertime heat flux to the underside of east Antarctic pack ice. J. Geophys. Res., 105 (C12), 28 75928 769.

    • Search Google Scholar
    • Export Citation
  • Manabe, S., and R. J. Stouffer, 1980: Sensitivity of a global climate model to an increase of CO2 concentration in the atmosphere. J. Geophys. Res., 85 (C10), 55295554.

    • Search Google Scholar
    • Export Citation
  • Martinson, D. G., 1990: Evolution of the Southern Ocean winter mixed layer and sea ice: Open ocean deepwater formation and ventilation. J. Geophys. Res., 95 (C7), 11 64111 654.

    • Search Google Scholar
    • Export Citation
  • Martinson, D. G., and R. A. Iannuzzi, 1998: Antarctic ocean-ice interaction: Implications from ocean bulk property distributions in the Weddell Gyre. Antarctic Sea Ice: Physical Processes, Interactions, and Variability, M. Jeffries, Ed., Antarctic Research Series, Vol. 74, Amer. Geophys. Union, 243–271.

    • Search Google Scholar
    • Export Citation
  • Middleton, J. H., and T. D. Foster, 1980: Fine structure measurements in a temperature-compensated halocline. J. Geophys. Res., 85 (C2), 11071122.

    • Search Google Scholar
    • Export Citation
  • Muench, R. D., H. J. S. Fernando, and G. R. Stegen, 1990: Temperature and salinity staircases in the northwestern Weddell Sea. J. Phys. Oceanogr., 20, 295306.

    • Search Google Scholar
    • Export Citation
  • Nelson, D. M., and W. O. Smith, 1991: Sverdrup revisited: Critical depths, maximum chlorophyll levels, and the control of Southern Ocean primary productivity by the irradiance-mixing regime. Limnol. Oceanogr., 36, 16501661.

    • Search Google Scholar
    • Export Citation
  • Orsi, A. H., T. Whitworth III, and W. D. Nowlin Jr., 1995: On the meridional extent and fronts of the Antarctic Circumpolar Current. Deep-Sea Res. I, 42, 641673.

    • Search Google Scholar
    • Export Citation
  • Owens, W. B., and A. P. S. Wong, 2009: An improved calibration method for the drift of the conductivity sensor on autonomous CTD profiling floats by θ-S climatology. Deep-Sea Res. I, 56, 450457.

    • Search Google Scholar
    • Export Citation
  • Riser, S. C., L. Ren, and A. Wong, 2008: Salinity in Argo: A modern view of a changing ocean. Oceanography, 21, 5667.

  • Ruddick, B., 1997: Differential fluxes of heat and salt: Implications for circulation and ecosystem modeling. Oceanography, 10, 122127.

    • Search Google Scholar
    • Export Citation
  • Schmitt, R. W., 1981: Form of the temperature-salinity relationship in the central water: Evidence for double-diffusive mixing. J. Phys. Oceanogr., 11, 10151026.

    • Search Google Scholar
    • Export Citation
  • Speer, K. G., and A. Forbes, 1994: A deep western boundary current in the South Indian Basin. Deep-Sea Res. I, 41, 12891303.

  • Spreen, G., L. Kaleschke, and G. Heygster, 2008: Sea ice remote sensing using AMSR-E 89-GHz channels. J. Geophys. Res., 113, C02S03, doi:10.1029/2005JC003384.

    • Search Google Scholar
    • Export Citation
  • Turner, J. S., 1973: Buoyancy Effects in Fluids. Cambridge University Press, 367 pp.

  • Whitworth, T., III, A. H. Orsi, and W. D. Nowlin Jr., 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, Ed., Antarctic Research Series, Vol. 75, Amer. Geophys. Union, 1–27.

    • Search Google Scholar
    • Export Citation
  • Williams, G. D., and N. L. Bindoff, 2003: Wintertime oceanography of the Adèlie Depression. Deep-Sea Res. II, 50, 13731392.

  • Worby, A. P., N. L. Bindoff, V. I. Lytle, I. Allison, and R. A. Massom, 1996: Winter ocean/sea ice interactions studied in the East Antarctic. Eos, Trans. Amer. Geophys. Union, 77, 453457.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    (a) Estimated trajectories of the 19 profiling floats used in this study. Triangles mark last position in each trajectory as of April 2010. The path of float with WMO ID 2900126 (Fig. 6) is highlighted in black. (b) Positions of the 1964 CTD profiles used in this study, collected between February 2007 and April 2010. The gray contour line marks the 2000-m isobath. The black dashed line marks the approximate position of the maximum winter sea ice edge. Schematic flow of ACC and ASC are also shown.

  • View in gallery

    False-color maps of (a) winter and (b) summer sea ice concentration in the Antarctic, produced by the Institute of Environmental Physics at the University of Bremen (Spreen et al. 2008). The oval shows the region containing the float data used in this study.

  • View in gallery

    The θS relations for water masses with σ0 > 27.5 kg m−3 from two regimes: (a) north of the ACC boundary and (b) south of the ACC boundary. Isopycnals σ0 = 27.5, 27.7, and 27.8 kg m−3 are also shown.

  • View in gallery

    Examples of thermohaline staircases in the permanent pycnocline in the Antarctic zone, where cold, fresh water overlies warm, salty water. The two CTD profiles are from the float with WMO ID 2900125.

  • View in gallery

    Percentage of under-ice profiles per month. This serves as a proxy of seasonal ice coverage in the area.

  • View in gallery

    Time series of measurements from the top 500 dbar from the float with WMO ID 2900126. Measurements were taken from October 2007 to April 2010. Missing data in the top 30 dbar indicate periods under sea ice. The white solid line marks approximate position of the mixed layer base. The black solid line is the 34.5 isohaline, which marks the approximate top of the permanent pycnocline. Brunt–Väisälä frequency N is defined as N2 = −(g/ρ)(/dz), where g is the acceleration of gravity.

  • View in gallery

    (a) Depth-averaged potential temperature in the mixed layer, (b) depth-averaged salinity in the mixed layer, and (c) mixed layer depths from profiles in water depths greater than 2000 m. Values are binned into the 12 months in the annual cycle. The solid line denotes mean value for each month.

  • View in gallery

    The amount of (a) entrainment Δhe, (b) mixed layer salinity change ΔS, and (c) ice growth Δhi, as a function of time over winter. Dashed (w = 0) and dotted (w = −we) lines denote amounts estimated by Eqs. (1)(3) for the Wilkes Land coast by using parameters listed in Table 2. Solid lines in (a) and (b) denote changes in mean mixed layer depth and salinity observed in float data from June to October (see Fig. 7).

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 319 214 25
PDF Downloads 191 117 16

Profiling Float Observations of the Upper Ocean under Sea Ice off the Wilkes Land Coast of Antarctica

View More View Less
  • 1 School of Oceanography, University of Washington, Seattle, Washington
Full access

Abstract

Multiyear under-ice temperature and salinity data collected by profiling floats are used to study the upper ocean near the Wilkes Land coast of Antarctica. The study region is in the seasonal sea ice zone near the southern terminus of the Antarctic Circumpolar Current. The profiling floats were equipped with an ice-avoidance algorithm and had a survival rate of 74% after 2.5 yr in the ocean. The data show that, in this part of Antarctica, the rate of sea ice decay exceeds the rate of sea ice growth. During the sea ice growth period, the water column is weakly stratified because of brine rejection and is only marginally stable. The average winter mixed layer temperature is about 0.12°C above the surface freezing point, providing evidence of entrainment of warmer water from the permanent pycnocline. The average mixed layer salinity increases by 0.127 from June to October. A one-dimensional model is used to quantify evolution of the winter mixed layer under a sea ice cover. The local winter entrainment rate is estimated to be 49 ± 11 m over 5 months, supplying a heat flux of 34 ± 8 W m−2 to the base of the mixed layer in winter. Model output gives a thermodynamic sea ice growth of 28 ± 15 cm over the same period. The winter ocean–atmosphere heat loss through leads and sea ice is estimated to be 14–25 W m−2 in this area, which is broadly in line with other winter observations from the East Antarctic region.

Corresponding author address: Dr. Annie Wong, School of Oceanography, University of Washington, Campus Box 355351, Seattle, WA 98195. E-mail: awong@ocean.washington.edu

Abstract

Multiyear under-ice temperature and salinity data collected by profiling floats are used to study the upper ocean near the Wilkes Land coast of Antarctica. The study region is in the seasonal sea ice zone near the southern terminus of the Antarctic Circumpolar Current. The profiling floats were equipped with an ice-avoidance algorithm and had a survival rate of 74% after 2.5 yr in the ocean. The data show that, in this part of Antarctica, the rate of sea ice decay exceeds the rate of sea ice growth. During the sea ice growth period, the water column is weakly stratified because of brine rejection and is only marginally stable. The average winter mixed layer temperature is about 0.12°C above the surface freezing point, providing evidence of entrainment of warmer water from the permanent pycnocline. The average mixed layer salinity increases by 0.127 from June to October. A one-dimensional model is used to quantify evolution of the winter mixed layer under a sea ice cover. The local winter entrainment rate is estimated to be 49 ± 11 m over 5 months, supplying a heat flux of 34 ± 8 W m−2 to the base of the mixed layer in winter. Model output gives a thermodynamic sea ice growth of 28 ± 15 cm over the same period. The winter ocean–atmosphere heat loss through leads and sea ice is estimated to be 14–25 W m−2 in this area, which is broadly in line with other winter observations from the East Antarctic region.

Corresponding author address: Dr. Annie Wong, School of Oceanography, University of Washington, Campus Box 355351, Seattle, WA 98195. E-mail: awong@ocean.washington.edu

1. Introduction

Each year, from early autumn to late spring, the heat lost from the sea surface around Antarctica leads to the development of a seasonal sea ice cover whose extent, at its maximum, essentially doubles the surface area of the continent. The presence of sea ice has a direct influence on the physical and biological processes in the ocean in the region. For example, the melting of sea ice in summer results in a stable and shallow surface mixed layer into which solar radiation can penetrate, thus providing conditions favorable for phytoplankton blooms (Nelson and Smith 1991). Over the broad continental shelves of the Weddell and Ross Seas, brine rejection associated with sea ice formation results in dense shelf waters, which mix with surrounding deep waters to form cold, dense, and highly oxygenated Antarctic Bottom Water (e.g., Gordon 1971; Gill 1973). Sea ice thickness, determined by the balance of heat and freshwater fluxes between the atmosphere, ice, and the ocean, affects surface albedo and is therefore an important contributor to the variability observed in global climate models (Manabe and Stouffer 1980). The study of the interaction between sea ice and the underlying ocean is therefore worthy of a considerable effort.

As sea ice forms, the salt in seawater is expelled into the ocean below, thus adding to the density of the underlying water mass. The sinking of this dense water mass and the compensating upwelling create a convective regime that transports salt from the surface to greater depths and entrains deeper waters into shallower levels. The deep-to-shallow oceanic heat flux associated with cross-pycnocline mixing is believed to be important in the balance of the annual sea-to-air heat transfer in the seasonal sea ice zone and in providing heat for the spring ice melt (Gordon 1981). Gordon et al. (1984) documented winter oceanographic observations east of the Greenwich meridian between 56° and 62°S, obtained during an expedition to the Weddell Sea in October and November 1981. Weddell Deep Water (WDW) entrainment into the winter mixed layer was inferred from the observed mixed layer oxygen deficit. The associated average heat flux during the 5-month ice-covered period was estimated to be 20–25 W m−2, with an effective WDW entrainment equivalent to 27 m of water, requiring an annual freshwater input of 34 cm into the surface layer to balance the WDW salt flux. Another winter expedition between July and September 1986 that surveyed a wider area allowed Gordon and Huber (1990) to estimate the winter heat flux for the 60°–70°S segment along the Greenwich meridian as 41 W m−2, with the average WDW entrainment equivalent to 45 m of water during the ice-covered period, requiring an annual compensating freshwater flux of 75 cm.

The advent of profiling float technology has enabled unprecedented year-round observation of the upper 2000 m of the global ocean. However, float observations from regions covered by sea ice have been sparse. Early attempts to observe the ice-covered ocean by profiling floats showed high instrument mortality rate, either because of crushing between ice floes while at the sea surface or hitting the bottom of the ice during ascent. Temperature criteria were subsequently explored as a means to detect the presence of sea ice and to instruct the float to abort its ascent to the sea surface. Such an ice-sensing algorithm was first used on profiling floats in the Weddell Sea in 2002, with instrument survival rate reported to be about 80% through the first winter in the ocean, more than a twofold increase from cases without the ice-sensing algorithm (Klatt et al. 2007).

In conjunction with the International Polar Year program, the University of Washington (UW) deployed 19 profiling floats programmed with an ice-avoidance algorithm along the East Antarctic coast between 80° and 130°E, an area historically referred to as the Wilkes Land coast (Fig. 1). These 19 profiling floats were deployed during two cruises, a February 2007 repeat of World Ocean Circulation Experiment line I8S from the RV Roger Revelle and a September–October 2007 expedition of the Australian icebreaker RSV Aurora Australis. As of April 2010, after having been in the ocean for three austral winters, 14 of these floats remained active, yielding a survival rate of 74% after 2.5 years. Other profiling floats equipped with the same ice-avoidance algorithm were also deployed in the Bellingshausen Sea and the Weddell Sea during 2007 and 2008.

Fig. 1.
Fig. 1.

(a) Estimated trajectories of the 19 profiling floats used in this study. Triangles mark last position in each trajectory as of April 2010. The path of float with WMO ID 2900126 (Fig. 6) is highlighted in black. (b) Positions of the 1964 CTD profiles used in this study, collected between February 2007 and April 2010. The gray contour line marks the 2000-m isobath. The black dashed line marks the approximate position of the maximum winter sea ice edge. Schematic flow of ACC and ASC are also shown.

Citation: Journal of Physical Oceanography 41, 6; 10.1175/2011JPO4516.1

This paper describes the data collected by the 19 profiling floats along the Wilkes Land coast from the time of their deployment through April of 2010 (Table 1). These data represent the first direct measurements of the annual cycle of the upper ocean under a sea ice cover in this region of the East Antarctic. In what follows, we focus on the processes that take place in the top 500 m of the upper ocean, including evolution of the mixed layer under a seasonal sea ice cover and the associated interaction with deeper waters. We show evidence of entrainment of warmer and saltier water from the permanent pycnocline into the winter mixed layer. The one-dimensional winter mixed layer model formulated by Martinson (1990, hereafter M90) is used to estimate entrainment rate, sea ice growth, and the heat and freshwater flux into the winter mixed layer.

Table 1.

The 19 profiling floats used in this study. Data are available from the Argo Global Data Centers (http://www.coriolis.eu.org; http://www.usgodae.org) by using WMO ID of the floats. Also shown are the UW identification numbers (UW ID).

Table 1.

2. Data description

The 19 profiling floats were programmed to park and drift with the flow at a pressure of 1000 dbar. At 7-day intervals, the floats descended to a pressure of 2000 dbar and then ascended and collected conductivity–temperature–depth (CTD) data to near the sea surface. The floats ascended at a nominal rate of about 6 cm s−1, with CTD data collected at a rate of 1 Hz. The CTD data were binned and averaged onboard each float and stored at 2-dbar intervals. The data were transmitted via the Iridium satellite system upon reaching the sea surface after completing each 7-day drift-and-rise cycle, unless surface sea ice was detected. The ice-avoidance algorithm used by these floats was modified slightly from that discussed by Klatt et al. (2007). Here, the determination was made that the float was under sea ice by calculating the median temperature between 50 and 30 dbar as the float ascended through the mixed layer. If this median temperature was below −1.78°C, the float was assumed to be profiling under sea ice. In such a case, the float stopped its ascent (usually at depths between 5 and 10 m) and descended back to 1000 dbar to begin the next 7-day cycle, without the transmission of any data. Instead, the data were stored onboard the float until the sea surface was judged by the ice-avoidance algorithm to be free of ice. There was enough memory on each float to collect and store as many as 68 CTD profiles in this fashion, although in practice it was usually not necessary to store more than about 40 profiles between contacts with the satellite. Data transmission through the Iridium system was sufficiently fast that each profile could be uploaded in no more than 2 min. The choice of −1.78°C as the cutoff temperature in the ice-avoidance algorithm followed Klatt et al. (2007), who determined that value by examining winter data available from the Weddell Sea.

Under-ice float tracking is possible in some parts of the Antarctic where an array of moored acoustic (RAFOS) sources exists (e.g., the Weddell Sea; see Klatt et al. 2007). However, there were no RAFOS sources in the Wilkes Land coast study region. Therefore, it was necessary to estimate the position of the CTD profiles by interpolation while the floats were under ice. To estimate the unknown position of the under-ice profiles, we used the known sampling time of the profiles, which was telemetered with the data, and linearly interpolated between the last known float position before sea ice developed and the first known float position after the spring thaw in the time series. Linearly interpolating between known positions is a simple way to estimate the mean direction of flow, but it does not account for eddy variability in the trajectories while the floats are under ice. For example, the mean distance between each consecutive pair of reported positions during the ice-free months is 32 km, whereas the mean distance between each consecutive pair of interpolated positions during the ice-covered months is only 13 km. Nonetheless, this simple method of estimating position under ice is sufficient for regional studies such as this one.

This study used 1964 quality-controlled CTD profiles that were located within the seasonal sea ice zone. Of the 1964 profiles, 1128 (57%) were collected under sea ice. Float salinity measurements were checked for conductivity sensor accuracy and precision using the statistical comparison method of Owens and Wong (2009) in conjunction with the World Ocean Database 2005. No evidence of conductivity sensor drift was detected in any of the 19 floats used in this study. Apart from six profiles that showed temporary effects of contamination (foreign matter likely entered the conductivity cell and was then expelled), no other salinity errors were detected. Float salinity accuracy was estimated to be better than 0.01 [practical salinity scale (PSS)-78], judging by the spread of salinity on isotherms below 500 dbar and from agreement with nearby historical data. Such a degree of float salinity accuracy was consistent with results reported in Riser et al. (2008). For temperature and pressure, the data appeared to be consistent with the manufacturer’s specifications of 0.002°C [International Temperature Scale of 1990 (ITS-90)] for temperature and 2.4 dbar for pressure.

3. Regional setting and water masses

Our study region lies between 70° and 130°E, south of 60°S, and is therefore south of the Southern Antarctic Circumpolar Current Front (SACCF; see Fig. 11 in Orsi et al. 1995). The study area lies in the open ocean and does not encompass the shallow continental shelf. Over 90% of the CTD profiles used were collected seaward of the 2000-m isobath, with the remainder collected at water depths between 1000 and 2000 m. The study region is covered by sea ice for 8–9 months of the year (Fig. 2).

Fig. 2.
Fig. 2.

False-color maps of (a) winter and (b) summer sea ice concentration in the Antarctic, produced by the Institute of Environmental Physics at the University of Bremen (Spreen et al. 2008). The oval shows the region containing the float data used in this study.

Citation: Journal of Physical Oceanography 41, 6; 10.1175/2011JPO4516.1

Circumpolar Deep Water (CDW), a water mass that is found exclusively in the Antarctic Circumpolar Current (ACC), rises toward the south near the Antarctic continent because of upwelling associated with the Antarctic Divergence and can be observed at depths as shallow as 100 m at its southern terminus. The water mass is generally further separated into Upper CDW (UCDW), characterized by relatively low oxygen and high nutrient concentrations, and Lower CDW (LCDW), characterized by relatively high salinities. The southern terminus of UCDW is commonly used to represent the poleward boundary of the ACC (Orsi et al. 1995). This is a transition zone with the eastward-flowing ACC to the north and the westward-flowing Antarctic Slope Current (ASC) to the south. Most of the floats migrated westward following the ASC. Near 85°E, one float [World Meteorological Organization buoy identification number (WMO ID) 4900478] migrated north of the transition zone (Fig. 1a), presumably joining a branch of the ACC that flows north along the eastern flank of the Kerguelen Plateau as a western boundary current (Speer and Forbes 1994).

Along the Wilkes Land coast, UCDW was observed to have potential density σ0 > 27.7 kg m−3, salinity S > 34.6, and potential temperature θ > 1.5°C (Fig. 3a). The denser LCDW, with σ0 > 27.8 kg m−3 and S > 34.7, penetrates farther south than the ACC boundary. The densest water mass in our dataset is Antarctic Bottom Water (AABW), with σ0 > 27.84 kg m−3 and θ < 0°C. Because the floats only profiled to 2000 m, AABW was only observed along the continental slope south of the ACC boundary (Fig. 3b).

Fig. 3.
Fig. 3.

The θS relations for water masses with σ0 > 27.5 kg m−3 from two regimes: (a) north of the ACC boundary and (b) south of the ACC boundary. Isopycnals σ0 = 27.5, 27.7, and 27.8 kg m−3 are also shown.

Citation: Journal of Physical Oceanography 41, 6; 10.1175/2011JPO4516.1

Lying above CDW is the permanent pycnocline. Water properties within the permanent pycnocline are generally saltier (and therefore denser) south of the ACC boundary than north of it, because of vertical mixing with the saltier LCDW (Fig. 3). The permanent pycnocline is the transition between CDW and Antarctic Surface Water (AASW). Properties of AASW are extremely variable, reflecting the variability at the air/sea/ice interface. At the sea surface in summer, AASW temperature can be as high as 3.0°C, with salinity as low as 32.5. Away from the sea surface, AASW temperatures and salinities are generally below 0°C and 34.5, respectively.

The permanent pycnocline in the Antarctic zone is therefore an environment where cold, fresh water overlies warmer, saltier water. Such an environment is conducive to double diffusion, where the relatively rapid molecular diffusion of temperature creates an instability, and therefore convection, in the region of highest vertical gradients, whereas the slower molecular diffusion of salt prevents the salt flux from becoming large (Ruddick 1997). The convection produces a quasi-homogeneous layer, above which another boundary is formed. This process migrates upward, resulting in a breakdown of the smooth stratification into a series of well-mixed layers separated by sharp interfaces or staircases. Such thermohaline staircases have previously been observed in the Weddell Sea in the transition from AASW to the underlying Warm Deep Water (Foster and Carmack 1976; Middleton and Foster 1980; Muench et al. 1990).

In our dataset, thermohaline staircases can be seen in the permanent pycnocline between depths of 200 and 400 m, with each step typically about 20 m thick (Fig. 4). To examine these steps in more detail, we calculated the dimensionless density ratio Rρ = βΔS/αΔT by following the definition of Huppert (1971) [note that this is the inverse of the definition of density ratio used by Schmitt (1981) in his study of Central Water]. Here, α is the thermal expansion coefficient, β is the haline contraction coefficient, and ΔS and ΔT are the characteristic vertical differences of salinity and temperature across the steps. By using this definition, the average density ratio over the depth range of 250–350 m is estimated to be 1.69 for the two CTD profiles shown in Fig. 4. This can be compared to the average values from the Weddell Sea of 1.03 by Foster and Carmack (1976), 1.4 by Middleton and Foster (1980), and 1.36 by Muench et al. (1990). Huppert (1971) has shown that staircases with 1 < Rρ < 2 are inherently unstable and will tend to break down in time. Hence, as with previous observations from the Weddell Sea, our dataset provides circumstantial evidence that mixing by diffusive instability plays some role in mixing of the water masses in the permanent pycnocline along the Wilkes Land coast.

Fig. 4.
Fig. 4.

Examples of thermohaline staircases in the permanent pycnocline in the Antarctic zone, where cold, fresh water overlies warm, salty water. The two CTD profiles are from the float with WMO ID 2900125.

Citation: Journal of Physical Oceanography 41, 6; 10.1175/2011JPO4516.1

4. Seasonal evolution of the mixed layer

Summer in Antarctica is short lived. The percentage of under-ice profiles over the course of a year (Fig. 5) indicates that, in this region, summer lasts only 3 months, from January to March. Most sea ice growth occurs in autumn during the months from April to June, with the ice cover remaining above 90% from July to October. Sea ice begins to retreat in November, with most of the spring thaw occurring in December. Thus, in this region, the rate of sea ice decay (occurring over 2 months) is greater than the rate of sea ice growth (which lasts 7 months).

Fig. 5.
Fig. 5.

Percentage of under-ice profiles per month. This serves as a proxy of seasonal ice coverage in the area.

Citation: Journal of Physical Oceanography 41, 6; 10.1175/2011JPO4516.1

The seasonal evolution of the mixed layer in the Antarctic zone is naturally related to the seasonal cycle of sea ice growth and decay. To illustrate the typical temporal evolution of the mixed layer under sea ice, we present the time series of measurements from the float with WMO ID 2900126 (Fig. 6). Here, we define the mixed layer base at each profile as the depth at which σ0 increases by 0.05 kg m−3 from the shallowest measurement.

Fig. 6.
Fig. 6.

Time series of measurements from the top 500 dbar from the float with WMO ID 2900126. Measurements were taken from October 2007 to April 2010. Missing data in the top 30 dbar indicate periods under sea ice. The white solid line marks approximate position of the mixed layer base. The black solid line is the 34.5 isohaline, which marks the approximate top of the permanent pycnocline. Brunt–Väisälä frequency N is defined as N2 = −(g/ρ)(/dz), where g is the acceleration of gravity.

Citation: Journal of Physical Oceanography 41, 6; 10.1175/2011JPO4516.1

a. Temperature

Mixed layer temperature remains almost uniform from June to October (Figs. 6a, 7a). This uniformity is due to the winter sea ice cover insulating the underlying ocean from any atmospheric forcing while providing a near-freezing temperature that is mixed downward by active brine rejection. This suggests that the process of brine rejection does not alter the temperature. The average mixed layer temperature between June and October is −1.76°C (Fig. 7a), about 0.12°C above the freezing point at the sea surface (assuming an average salinity of 34.27). The slightly warmer winter mixed layer temperature relative to the surface freezing point is the result of incorporation of water from the permanent pycnocline. Although the winter mixed layer is continually cooled from above by sea ice, it is also warmed from below by pycnocline water, which derives its heat from CDW. The temperature of water underlying the base of the winter mixed layer can reach 0°C by October, when the winter mixed layer is deepest. The lowest temperature recorded in our dataset is −1.883°C at approximately 30 dbar in August, with salinity of 34.34.

Fig. 7.
Fig. 7.

(a) Depth-averaged potential temperature in the mixed layer, (b) depth-averaged salinity in the mixed layer, and (c) mixed layer depths from profiles in water depths greater than 2000 m. Values are binned into the 12 months in the annual cycle. The solid line denotes mean value for each month.

Citation: Journal of Physical Oceanography 41, 6; 10.1175/2011JPO4516.1

The mixed layer shoals rapidly during November and December and remains shallow through early March. At the height of the Antarctic summer, when the ice-free ocean is under direct influence of the atmosphere, the shallow mixed layer can be warmed to near 3°C (Fig. 7a). The remnant signal from the previous winter can be seen as a subsurface temperature minimum, often referred to as Winter Water, which resides between the relatively warm summer mixed layer above and the permanent pycnocline below (Fig. 6a). This summer temperature minimum retains the near-freezing characteristic of the winter mixed layer but is eroded away by early April, as the mixed layer begins to deepen at the onset of winter. AASW (Fig. 3), often defined as the water mass that extends from the sea surface to the base of the mixed layer in autumn and winter or the base of Winter Water in spring and summer (Whitworth et al. 1998), can be effectively separated from the permanent pycnocline in this region by the 34.5 isohaline (Fig. 6).

b. Salinity, density, and stability

The melting of sea ice creates a fresh and shallow summer mixed layer and a shallow halocline during the period from January to March (Fig. 6b). This pool of freshwater represents a salt deficit that is stored in the mixed layer until the seasonal halocline is eroded in April and May, when initial sea ice growth begins to eliminate the summer deficit via brine rejection. As the mixed layer deepens from April to October, a series of increasingly salty isohaline columns appear (Fig. 6b), providing evidence of convective mixing due to the brine rejected from active sea ice formation. The increase in mixed layer salinity over the course of autumn and winter is therefore partly the result of accumulation of expelled salt and partly due to entrainment of saltier pycnocline water. The average mixed layer salinity reaches a maximum of 34.31 in October (Fig. 7b), when the winter mixed layer is deepest. The high winter mixed layer salinity persists into summer at the base of Winter Water (Fig. 6b), to be recycled into the mixed layer when sea ice growth resumes.

Because temperature only varies over a small range at these high latitudes, the density distribution in the mixed layer is largely determined by salinity distribution (Fig. 6c). The average mixed layer density σθ can be as low as 26.0 kg m−3 in summer and as high as 27.7 kg m−3 in winter. In summer, the shallow halocline formed from melting ice increases the stratification and stability of the AASW cap at the sea surface. This increased stability facilitates freezing later in the seasonal cycle as winter approaches. In winter, the mixed layer is very weakly stratified because of brine rejection and is only marginally stable (Fig. 6d).

Stronger stratification and therefore increased stability are seen along the base of the winter mixed layer from April to October (Fig. 6d). This subsurface stratified layer represents the permanent pycnocline, a zone where density changes rapidly from winter mixed layer water to CDW. Although the presence of CDW contributes stronger vertical density gradients, the Southern Ocean pycnocline south of the Polar Front is, in general, weaker than the permanent pycnocline of other oceans because of the small freshwater input into the surface layer (Gordon 1981).

The depth of winter convective mixing in the Antarctic zone is governed by the magnitude of external forcing and the relative strength of the permanent pycnocline. The strength of the permanent pycnocline is, in turn, controlled by the presence or absence of CDW. Where CDW is present, a relatively strong pycnocline ensues, and the resulting enhanced subsurface stratification acts as a barrier to convection. In the open ocean where CDW is present, winter mixed layer depths generally do not exceed about 300 m (Fig. 7c). Where CDW is absent and subsurface stratification is weaker, as is the case on the continental slope and shelf, convection associated with brine rejection can drive the winter mixed layer to greater depths than those shown in Fig. 7c. One float (WMO ID 2900118) wandered onto an area of the continental slope where CDW was absent and recorded mixed layer depths of about 500 m during June 2008 (not shown).

5. Heat and freshwater flux in the winter mixed layer

The intricate relationships between convective mixing from sea ice growth, deep-water entrainment, and evolution of the winter mixed layer have been studied by M90 through the development of an analytical winter mixed layer model. A fundamental assumption of the M90 model is that the ocean–atmosphere–ice system is controlled predominantly by vertical processes. This assumption is reasonable in regions where lateral property gradients are weak, as in the open ocean away from fronts, topographic features, and western boundaries. As such, this winter mixed layer model is one dimensional in the vertical and involves two explicit assumptions. First, entrainment of deep water into the winter mixed layer is driven mainly by convection induced by brine rejection during ice growth, whereas turbulence produced by relative ice motion and/or wind in winter serves to maintain the well-mixed surface layer through winter. This assumption downplays the role of entrainment due to turbulence-induced mixing, which is reasonable because buoyancy forcing is the dominant factor in the winter entrainment rate. Second, the winter mixed layer heat storage is ignored, because it influences the system over relatively short time scales. Any heat input into the winter mixed layer is lost to the atmosphere through the sea ice cover and leads, or it is used to melt the underside of the ice.

The model identifies three dominant vertical processes that control the delicate balance in the upper water column under a sea ice cover. These are 1) heat loss from the winter mixed layer to the atmosphere by conduction through the ice and directly through the leads; 2) heat and salt flux across the pycnocline by entrainment and diffusive mixing; and 3) ice growth, which introduces a salt flux. As functions of a set of oceanic conditions, M90 found solutions for entrainment depth Δhe, mixed layer salinity change ΔS, and sea ice growth Δhi to be
e1
e2
e3

Equations (1)(3) describe the winter evolution of the water column under a sea ice cover as a function of time t (s) that starts after erosion of the seasonal pycnocline in autumn. The parameters used in these equations are explained in details in M90 and are summarized in Tables 2 and 3. When used with winter observations from the Weddell Sea, simulations from M90 have been shown to agree well with descriptive analysis results (e.g., Gordon and Huber 1990). Here, we use the M90 model and the winter upper-ocean conditions observed by the profiling floats to estimate winter heat loss from ocean to atmosphere under a sea ice cover and sea ice production in the Wilkes Land coast region. This region does not have significant lateral property gradients (Fig. 6) and therefore matches the general conditions assumed by the M90 model.

Table 2.

Parameter values and constants used to generate Fig. 8.

Table 2.
Table 3.

Model terms derived from the parameters and constants in Table 2.

Table 3.

a. Model results and sensitivity

The three main ocean interior parameters in the M90 model are the initial winter mixed layer depth h0 and temperature and salinity gradients across the permanent pycnocline, T and S. A strong temperature gradient across the pycnocline limits ice growth, whereas the amount of entrainment is inversely proportional to the pycnocline salinity gradient. The depth of the mixed layer controls the volume that is influenced by external forcing and therefore controls the amount of ice growth. We assign h0 = 90 m, which is the observed average mixed layer depth in June (Fig. 7c). The values of T and S are computed as the spatially averaged pycnocline gradients of all profiles from June to October. The unknown heat loss parameter Fatm, representing winter heat loss from ocean to atmosphere through leads and snow-covered ice, is obtained by adjustment within an estimated range in order to provide model results for ΔS [Eq. (2)] that are consistent with the observed mean depth-averaged mixed layer salinity increase of 0.127 from June to October (Fig. 7b). Other constants used are the same as in M90. The model was run for 150 days from June to October, covering the period after fall erosion of the seasonal pycnocline and when the winter mixed layer is deepest.

We consider two extreme cases in the model by using the set of parameters listed in Table 2. The first case assumes there is no Ekman upwelling (w = 0; dashed lines in Fig. 8). With no upwelling, the absolute depth of the winter mixed layer is controlled by entrainment alone. The second case assumes that upwelling balances entrainment (w = −we; dotted lines in Fig. 8), so there is no net increase in absolute mixed layer depth. The amount of actual winter mixed layer deepening will lie somewhere between these two extreme cases, as shown in M90. Variations resulting from these two limiting cases amount to about 7% in winter entrainment Δhe and winter mixed layer salinity change ΔS, but about 30% in sea ice growth Δhi. The model overestimates the amount of entrainment but underestimates the amount of salt in the mixed layer in the middle part of winter (Fig. 8). This is presumably due to the treatment of Fatm as a constant. M90 has suggested that a time-varying Fatm might introduce more curvature to the solutions.

Fig. 8.
Fig. 8.

The amount of (a) entrainment Δhe, (b) mixed layer salinity change ΔS, and (c) ice growth Δhi, as a function of time over winter. Dashed (w = 0) and dotted (w = −we) lines denote amounts estimated by Eqs. (1)(3) for the Wilkes Land coast by using parameters listed in Table 2. Solid lines in (a) and (b) denote changes in mean mixed layer depth and salinity observed in float data from June to October (see Fig. 7).

Citation: Journal of Physical Oceanography 41, 6; 10.1175/2011JPO4516.1

The model sea ice growth estimate Δhi is sensitive to changes in the initial winter mixed layer depth h0, whereas entrainment Δhe is sensitive to pycnocline salinity gradient S and heat to salt ratio in the pycnocline rTS. The sensitivity of model output to changes in these parameters has been discussed in details in M90. Here, we attempt to estimate the uncertainty in Fatm, Δhe, and Δhi by forcing the model with a range of observed values for h0, S, and rTS while keeping the salinity difference ΔS = 0.127 fixed. The observed June average mixed layer depth of h0 = 90 m has a standard deviation of 30 m. When pycnocline gradients are kept fixed and h0 is allowed to vary within ±30 m, Table 4a shows that model sea ice growth estimates range from 13 to 43 cm, whereas model entrainment estimates remain constant at about 50 m. On the other hand, when h0 = 90 m is kept fixed and pycnocline gradients are allowed to change by varying the pycnocline thickness by ±30 m, model entrainment estimates range from 38 to 60 m, whereas model sea ice growth estimates remain in a relatively constant range of 22–36 cm (Table 4b). Therefore, altering the pycnocline gradients will require the entrainment rate to change to entrain a constant amount of salt.

Table 4.

Sensitivity of model-estimated winter entrainment Δhe and sea ice growth Δhi to (a) changes in initial winter mixed layer depth h0 and (b) changes in pycnocline salinity gradient S and ratio of heat to salt across the pycnocline rTS. The value of Fatm is obtained by fitting to ΔS = 0.127. Results are obtained after a model run of 150 days and are bounded by two extreme cases: w = 0 and w = −we. Other constants used are as in Table 2.

Table 4.

The amount of winter entrainment Δhe is therefore estimated to be 49 ± 11 m over 5 months, whereas sea ice growth Δhi is estimated to be 28 ± 15 cm over the same period. Combining 49 m of winter entrainment with the observed average June mixed layer depth of h0 = 90 m gives a model-estimated mixed layer depth of 139 m in October, similar to the observed October average of 137 m (Fig. 7c). The model estimate of thermodynamic sea ice growth of 28 cm over 5 months amounts to an average sea ice growth of about 6 cm month−1 over winter. This agrees with reports from a winter experiment conducted near 64°S, 140°E in August 1995, during which thermodynamic ice formation of up to 5 cm was observed at the base of existing floes (Worby et al. 1996).

The winter heat loss from ocean to the atmosphere through ice Fatm is estimated to be in the range of 19–25 W m−2 if double diffusion is assumed to be operative (rd = 3.3; Turner 1973). If double diffusion is ignored (rd = 1), Fatm is estimated to be in the range of 14–20 W m−2. Lytle et al. (2000) estimated the average winter heat flux from ocean to ice to be 13.0–14.5 W m−2 in the open ocean near 65°S, 140°E (see also Worby et al. 1996). They used two methods of estimation: a simple thermodynamic model that involved estimating the temperature gradient measured directly through the sea ice, and a turbulent heat flux model that relied on estimating ice and water velocities from drifting buoy data. Farther to the east, Williams and Bindoff (2003) used a mass conservation model and estimated an average winter ocean sensible heat flux to the atmosphere of 30 W m−2 near the Mertz polynya over the continental shelf between 143° and 146°E. Our ocean–atmosphere heat flux estimate is bracketed by these two previous winter studies and illustrates the sensitivity of ocean heat flux values to the methodology used in the estimation.

b. Heat and freshwater flux estimates

The heat per unit area Q associated with upward entrainment of water from the permanent pycnocline can be estimated as
e4
where ΔT = TpycTf. Assuming an average pycnocline temperature Tpyc = 0.25°C and using the freezing point of seawater Tf = −1.88°C at the sea surface yields Q = 43.6 ± 9.8 × 107 J m−2. Dividing this by the 5-month winter entrainment period yields an average entrainment heat flux estimate of 34 ± 8 W m−2 over this period, which is similar to the range of comparable winter estimates from the Weddell Sea (20–25 W m−2 from Gordon et al. 1984; 41 W m−2 from Gordon and Huber 1990).
Over the course of a year, the salt introduction to the mixed layer by winter entrainment is balanced by annual surface freshwater input in the region. The amount of freshwater H required to balance the salt flux associated with winter entrainment can be estimated by
e5
Assuming an average pycnocline salinity Spyc = 34.55, the entrainment of 49 ± 11 m of pycnocline water over the five ice-covered months requires an annual freshwater input H = 73 ± 16 cm to achieve the annual mean mixed layer salinity of Sml = 34.04. The annual freshwater input into the mixed layer can be due to excess precipitation over evaporation, glacial ice melt, and net sea ice advection into the region in addition to local sea ice melting. Estimates of freshwater input to the Southern Ocean are subject to large uncertainty. Nonetheless, our estimate of 73 ± 16 cm of annual freshwater input requirement for the Wilkes Land coast is similar to the 75 cm estimated by Gordon and Huber (1990) for the sea ice zone along the Greenwich meridian.

6. Conclusions and discussion

The annual cycle of the upper ocean along the Wilkes Land coast has been studied by using data from profiling floats equipped with an ice-avoidance algorithm. In the sea ice zone around Antarctica, the seasonal pycnocline is eroded by brine rejection during autumn, after which the mixed layer continues to deepen from June to October as a result of entrainment of water from the permanent pycnocline. By using the analytical model of M90, winter entrainment in the study region is estimated to be 49 ± 11 m over 5 months, with thermodynamic sea ice growth estimated to be 28 ± 15 cm over the same period. The average winter entrainment heat flux at the base of the mixed layer is estimated to be 34 ± 8 W m−2 over 5 months, comparable to similar winter estimates from the Weddell Sea. It is also estimated that an annual surface freshwater input of 73 ± 16 cm is needed to balance the salt flux in the mixed layer associated with winter entrainment. A future extension of this study is an analysis that will characterize more carefully the spatial and temporal distribution of heat and freshwater flux in the sea ice zone, in the manner of Martinson and Iannuzzi (1998).

We also demonstrate that a quantitative winter mixed layer model can provide a means to use in situ hydrographic data to estimate flux terms such as Fatm, the winter ocean heat loss to the atmosphere in the sea ice zone. These air–sea flux variables are traditionally difficult to estimate, especially in challenging environments such as the seasonal ice zone around Antarctica. Any basic advances in improving such estimates will help to improve the representations of these quantities in global climate models. We estimate the winter ocean–atmosphere heat loss in the region to be 14–25 W m−2, which is consistent with previous winter studies from the East Antarctic region (Lytle et al. 2000; Williams and Bindoff 2003).

The Wilkes Land coast, an open ocean region where the presence of CDW results in a warm and salty permanent pycnocline, may be representative of the sea ice zone in many locations around Antarctica. These are environments where entrainment can introduce heat and salt into the winter mixed layer. Entrainment heat flux drives a negative feedback: large entrainment will erode the pycnocline, allowing open ocean convection via brine rejection, whereas the upward heat flux associated with such convection can melt the sea ice cover, leading to polynya formation. Similar studies around Antarctica can provide insights into potential likely spots for the formation of these sensible heat polynyas and therefore deep water ventilation.

Our dataset has confirmed several observations made by Gordon (1981) regarding differences between the Arctic and the Antarctic. Sea ice decay is faster than sea ice growth around Antarctica, whereas the opposite is generally true in the Arctic. This contrast suggests fundamental differences in the temperature and salinity characteristics in the two polar oceans. The permanent pycnocline in the Southern Ocean is generally weaker than that in the Arctic Ocean, because of the relatively low freshwater input (and thus relatively high salinity) of surface water around Antarctica. Gordon (1981) noted that the rapid sea ice decay around Antarctica could partly be explained by the relatively weak pycnocline, which allowed easier vertical heat flux from pycnocline water into the surface water through Ekman pumping and cross-pycnocline mixing. The oceanic heat budget in the Arctic is presumably quite different from the winter model presented in this study. Our results suggest that observations from profiling floats that collect CTD data throughout the annual cycle, when the polar oceans are ice covered and when they are not, can be an important tool in understanding the interactions of the ocean, atmosphere, and sea ice in the high-latitude regions.

Acknowledgments

The authors wish to acknowledge the officers, crew, and scientists of the RV Roger Revelle during the CLIVAR/CO2 Repeat Hydrography I8S 2007 cruise and the RSV Aurora Australis during the SIPEX 2007 cruise for their assistance in deploying the profiling floats used in this study. Special thanks are given to Dr. Doug Martinson of the Lamont-Doherty Earth Observatory for comments relating to his winter mixed layer model. Comments from two anonymous reviewers helped improve the manuscript. The profiling floats used in this study were fabricated and programmed by Dana Swift, Dale Ripley, and Rick Rupan of the University of Washington. Robert Drucker helped with the visualization of the satellite sea ice extent data. The data used in this study were collected and made freely available as part of the International Argo Project and the national programs that contribute to it (http://www.argo.net). This work was funded by National Oceanic and Atmospheric Administration (NOAA) Grant NA17RJ1232 Task 2 to the University of Washington in support of the Argo float program.

REFERENCES

  • Foster, T. D., and E. C. Carmack, 1976: Temperature and salinity structure in the Weddell Sea. J. Phys. Oceanogr., 6, 3644.

  • Gill, A. E., 1973: Circulation and bottom water production in the Weddell Sea. Deep-Sea Res., 20, 111140.

  • Gordon, A. L., 1971: Oceanography of Antarctic waters. Antarctic Oceanology, Vol. 1, J. L. Reid, Ed., Antarctic Research Series, Vol. 15, Amer. Geophys. Union, 169–203.

    • Search Google Scholar
    • Export Citation
  • Gordon, A. L., 1981: Seasonality of Southern Ocean sea ice. J. Geophys. Res., 86 (C5), 41934197.

  • Gordon, A. L., and B. A. Huber, 1990: Southern Ocean winter mixed layer. J. Geophys. Res., 95 (C7), 11 65511 672.

  • Gordon, A. L., C. T. A. Chen, and W. G. Metcalf, 1984: Winter mixed layer entrainment of Weddell Deep Water. J. Geophys. Res., 89 (C1), 637640.

    • Search Google Scholar
    • Export Citation
  • Huppert, H. E., 1971: On the stability of a series of double-diffusive layers. Deep-Sea Res., 18, 10051021.

  • Klatt, O., O. Boebel, and E. Fahrbach, 2007: A profiling float’s sense of ice. J. Atmos. Oceanic Technol., 24, 13011308.

  • Lytle, V. I., R. Massom, N. Bindoff, A. Worby, and I. Allison, 2000: Wintertime heat flux to the underside of east Antarctic pack ice. J. Geophys. Res., 105 (C12), 28 75928 769.

    • Search Google Scholar
    • Export Citation
  • Manabe, S., and R. J. Stouffer, 1980: Sensitivity of a global climate model to an increase of CO2 concentration in the atmosphere. J. Geophys. Res., 85 (C10), 55295554.

    • Search Google Scholar
    • Export Citation
  • Martinson, D. G., 1990: Evolution of the Southern Ocean winter mixed layer and sea ice: Open ocean deepwater formation and ventilation. J. Geophys. Res., 95 (C7), 11 64111 654.

    • Search Google Scholar
    • Export Citation
  • Martinson, D. G., and R. A. Iannuzzi, 1998: Antarctic ocean-ice interaction: Implications from ocean bulk property distributions in the Weddell Gyre. Antarctic Sea Ice: Physical Processes, Interactions, and Variability, M. Jeffries, Ed., Antarctic Research Series, Vol. 74, Amer. Geophys. Union, 243–271.

    • Search Google Scholar
    • Export Citation
  • Middleton, J. H., and T. D. Foster, 1980: Fine structure measurements in a temperature-compensated halocline. J. Geophys. Res., 85 (C2), 11071122.

    • Search Google Scholar
    • Export Citation
  • Muench, R. D., H. J. S. Fernando, and G. R. Stegen, 1990: Temperature and salinity staircases in the northwestern Weddell Sea. J. Phys. Oceanogr., 20, 295306.

    • Search Google Scholar
    • Export Citation
  • Nelson, D. M., and W. O. Smith, 1991: Sverdrup revisited: Critical depths, maximum chlorophyll levels, and the control of Southern Ocean primary productivity by the irradiance-mixing regime. Limnol. Oceanogr., 36, 16501661.

    • Search Google Scholar
    • Export Citation
  • Orsi, A. H., T. Whitworth III, and W. D. Nowlin Jr., 1995: On the meridional extent and fronts of the Antarctic Circumpolar Current. Deep-Sea Res. I, 42, 641673.

    • Search Google Scholar
    • Export Citation
  • Owens, W. B., and A. P. S. Wong, 2009: An improved calibration method for the drift of the conductivity sensor on autonomous CTD profiling floats by θ-S climatology. Deep-Sea Res. I, 56, 450457.

    • Search Google Scholar
    • Export Citation
  • Riser, S. C., L. Ren, and A. Wong, 2008: Salinity in Argo: A modern view of a changing ocean. Oceanography, 21, 5667.

  • Ruddick, B., 1997: Differential fluxes of heat and salt: Implications for circulation and ecosystem modeling. Oceanography, 10, 122127.

    • Search Google Scholar
    • Export Citation
  • Schmitt, R. W., 1981: Form of the temperature-salinity relationship in the central water: Evidence for double-diffusive mixing. J. Phys. Oceanogr., 11, 10151026.

    • Search Google Scholar
    • Export Citation
  • Speer, K. G., and A. Forbes, 1994: A deep western boundary current in the South Indian Basin. Deep-Sea Res. I, 41, 12891303.

  • Spreen, G., L. Kaleschke, and G. Heygster, 2008: Sea ice remote sensing using AMSR-E 89-GHz channels. J. Geophys. Res., 113, C02S03, doi:10.1029/2005JC003384.

    • Search Google Scholar
    • Export Citation
  • Turner, J. S., 1973: Buoyancy Effects in Fluids. Cambridge University Press, 367 pp.

  • Whitworth, T., III, A. H. Orsi, and W. D. Nowlin Jr., 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, Ed., Antarctic Research Series, Vol. 75, Amer. Geophys. Union, 1–27.

    • Search Google Scholar
    • Export Citation
  • Williams, G. D., and N. L. Bindoff, 2003: Wintertime oceanography of the Adèlie Depression. Deep-Sea Res. II, 50, 13731392.

  • Worby, A. P., N. L. Bindoff, V. I. Lytle, I. Allison, and R. A. Massom, 1996: Winter ocean/sea ice interactions studied in the East Antarctic. Eos, Trans. Amer. Geophys. Union, 77, 453457.

    • Search Google Scholar
    • Export Citation
Save