• Barnston, A. G., , and R. E. Livezey, 1987: Classification, seasonality and persistence of low-frequency atmospheric circulation patterns. Mon. Wea. Rev., 115, 10831126, doi:10.1175/1520-0493(1987)115<1083:CSAPOL>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Bingham, F. M., , G. R. Foltz, , and M. J. McPhaden, 2012: Characteristics of the seasonal cycle of surface layer salinity in the global ocean. Ocean Sci., 8, 915929, doi:10.5194/os-8-915-2012.

    • Search Google Scholar
    • Export Citation
  • Bingham, F. M., , J. Busecke, , A. L. Gordon, , C. F. Giulivi, , and Z. Li, 2014: The North Atlantic subtropical surface salinity maximum as observed by Aquarius. J. Geophys. Res. Oceans, 119, 77417755, doi:10.1002/2014JC009825.

    • Search Google Scholar
    • Export Citation
  • Blanke, B., , M. Arhan, , A. Lazar, , J. Mignot, , and G. Prévost, 2002: A Lagrangian numerical investigation of the origins and fates of the salinity maximum water in the Atlantic. J. Geophys. Res., 107, 3163, doi:10.1029/2002JC001318.

    • Search Google Scholar
    • Export Citation
  • Boyer, T. P., , S. Levitus, , J. Antonov, , R. Locarnini, , and H. Garcia, 2005: Linear trends in salinity for the World Ocean, 1955–1998. Geophys. Res. Lett., 32, L01604, doi:10.1029/2004GL021791.

    • Search Google Scholar
    • Export Citation
  • Burkholder, K. C., , and M. S. Lozier, 2011: Subtropical to subpolar pathways in the North Atlantic: Deductions from Lagrangian trajectories. J. Geophys. Res., 116, C07017, doi:10.1029/2010JC006697.

    • Search Google Scholar
    • Export Citation
  • Busecke, J., , Z. Li, , F. M. Bingham, , A. L. Gordon, , and J. Font, 2014: Mesoscale turbulence within the subtropical North Atlantic surface layer. J. Geophys. Res. Oceans, 119, 41244140, doi:10.1002/2013JC009715.

    • Search Google Scholar
    • Export Citation
  • Chen, J., , T. Qu, , Y. Sasaki, , and N. Schneider, 2010: Anti-correlated variability in subduction rate of the western and eastern North Pacific Oceans as seen from an eddy-resolving ocean GCM. Geophys. Res. Lett., 37, L23608, doi:10.1029/2010GL045239.

    • Search Google Scholar
    • Export Citation
  • D’Addezio, J. M., , and F. M. Bingham, 2014: A subtropical North Atlantic regional atmospheric moisture budget. J. Geophys. Res. Oceans, 119, 87318748, doi:10.1002/2014JC010300.

    • Search Google Scholar
    • Export Citation
  • de Boer, A. M., , J. R. Toggweiler, , and D. M. Sigman, 2008: Atlantic dominance of the meridional overturning circulation. J. Phys. Oceanogr., 38, 435450, doi:10.1175/2007JPO3731.1.

    • Search Google Scholar
    • Export Citation
  • Deser, C., , and M. S. Timlin, 1997: Atmosphere–ocean interaction on weekly timescales in the North Atlantic and Pacific. J. Climate, 10, 393408, doi:10.1175/1520-0442(1997)010<0393:AOIOWT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Durack, P., , and S. Wijffels, 2010: Fifty-year trends in global ocean salinities and their relationship to broad-scale warming. J. Climate, 23, 43424362, doi:10.1175/2010JCLI3377.1.

    • Search Google Scholar
    • Export Citation
  • Giannini, A., , J. C. H. Chiang, , M. A. Cane, , Y. Kushnir, , and R. Seager, 2001: The ENSO teleconnection to the tropical Atlantic Ocean: Contributions of the remote and local SSTs to rainfall variability in the tropical Americas. J. Climate, 14, 45304544, doi:10.1175/1520-0442(2001)014<4530:TETTTT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Gordon, A., , and C. F. Giulivi, 2008: Sea surface salinity trends. Oceanography, 21, 2029, doi:10.5670/oceanog.2008.64.

  • Gu, D., , and S. G. H. Philander, 1997: Interdecadal climate fluctuations that depend on exchanges between the tropics and extratropics. Science, 275, 805807, doi:10.1126/science.275.5301.805.

    • Search Google Scholar
    • Export Citation
  • Häkkinen, S., 1999: Variability of the simulated meridional heat transport in the North Atlantic for the period 1951–1993. J. Geophys. Res., 104, 10 99111 008, doi:10.1029/1999JC900034.

    • Search Google Scholar
    • Export Citation
  • Held, I. M., , and B. J. Soden, 2006: Robust responses of the hydrological cycle to global warming. J. Climate, 19, 56865699, doi:10.1175/JCLI3990.1.

    • Search Google Scholar
    • Export Citation
  • Hormann, V., , L. R. Centurioni, , and G. Reverdin, 2015: Evaluation of drifter salinities in the subtropical North Atlantic. J. Atmos. Oceanic Technol., 32, 185192, doi:10.1175/JTECH-D-14-00179.1.

    • Search Google Scholar
    • Export Citation
  • IPCC, 2007: Climate Change 2007: The Physical Science Basis. Cambridge University Press, 996 pp.

  • Kalnay, E., and et al. , 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77, 437371, doi:10.1175/1520-0477(1996)077<0437:TNYRP>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Karstensen, J., , and D. Quadfasel, 2002: Water subducted into the Indian Ocean subtropical gyre. Deep-Sea Res. II, 49, 14411457, doi:10.1016/S0967-0645(01)00160-6.

    • Search Google Scholar
    • Export Citation
  • Kilpatrick, T., , N. Schneider, , and E. Di Lorenzo, 2011: Generation of low-frequency spiciness variability in the thermocline. J. Phys. Oceanogr., 41, 365377, doi:10.1175/2010JPO4443.1.

    • Search Google Scholar
    • Export Citation
  • Kolodziejczyk, N., , and F. Gaillard, 2012: Observation of spiciness interannual variability in the Pacific pycnocline. J. Geophys. Res., 117, C12018, doi:10.1029/2012JC008365.

    • Search Google Scholar
    • Export Citation
  • Kolodziejczyk, N., , G. Reverdin, , and A. Lazar, 2015: Interannual variability of the mixed layer winter convection and spice injection in the eastern subtropical North Atlantic. J. Phys. Oceanogr., 45, 504525, doi:10.1175/JPO-D-14-0042.1.

    • Search Google Scholar
    • Export Citation
  • Laurian, A., , A. Lazar, , G. Reverdin, , K. Rodgers, , and P. Terray, 2006: Poleward propagation of spiciness anomalies in the North Atlantic Ocean. Geophys. Res. Lett., 33, L13603, doi:10.1029/2006GL026155.

    • Search Google Scholar
    • Export Citation
  • Laurian, A., , A. Lazar, , and G. Reverdin, 2009: Generation mechanism of spiciness anomalies: An OGCM analysis in the North Atlantic Subtropical Gyre. J. Phys. Oceanogr., 39, 10031018, doi:10.1175/2008JPO3896.1.

    • Search Google Scholar
    • Export Citation
  • Lazar, A., , R. Murtugudde, , and A. J. Busalacchi, 2001: A model study of temperature anomaly propagation from subtropics to tropics within the South Atlantic thermocline. Geophys. Res. Lett., 28, 12711274, doi:10.1029/2000GL011418.

    • Search Google Scholar
    • Export Citation
  • Levitus, S., 1982: Climatological Atlas of the World Ocean. NOAA Prof. Paper 13, 173 pp. and 17 microfiche.

  • Liu, L. L., , and R. X. Huang, 2012: The global subduction/obduction rates: Their interannual and decadal variability. J. Climate, 25, 10961115, doi:10.1175/2011JCLI4228.1.

    • Search Google Scholar
    • Export Citation
  • Luo, Y., , L. M. Rothstein, , R.-H. Zhang, , and A. J. Busalacchi, 2005: On the connection between South Pacific subtropical spiciness anomalies and decadal equatorial variability in an ocean general circulation model. J. Geophys. Res., 110, C10002, doi:10.1029/2004JC002655.

    • Search Google Scholar
    • Export Citation
  • Mignot, J., , and C. Frankignoul, 2004: Interannual to interdecadal variability of sea surface salinity in the Atlantic and its link to the atmosphere in a coupled model. J. Geophys. Res., 109, C04005, doi:10.1029/2003JC002005.

    • Search Google Scholar
    • Export Citation
  • Niiler, P., 1975: Deepening of the wind-mixed layer. J. Mar. Res., 33, 405422.

  • O’Connor, B. M., , R. A. Fine, , and D. B. Olson, 2005: A global comparison of subtropical underwater formation rate. Deep-Sea Res., 52, 15691590, doi:10.1016/j.dsr.2005.01.011.

    • Search Google Scholar
    • Export Citation
  • Qiu, B., , and R. X. Huang, 1995: Ventilation of the North Atlantic and North Pacific: Subduction versus obduction. J. Phys. Oceanogr., 25, 23742390, doi:10.1175/1520-0485(1995)025<2374:VOTNAA>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Qu, T., , and J. Chen, 2009: A North Pacific decadal variability in subduction rate. Geophys. Res. Lett., 36, L22602, doi:10.1029/2009GL040914.

    • Search Google Scholar
    • Export Citation
  • Qu, T., , H. Mitsudera, , and T. Yamagata, 1999: A climatology of the circulation and water mass distribution near the Philippine coast. J. Phys. Oceanogr., 29, 14881505, doi:10.1175/1520-0485(1999)029<1488:ACOTCA>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Qu, T., , S. Gao, , and I. Fukumori, 2011: What governs the North Atlantic salinity maximum in a global GCM? Geophys. Res. Lett., 38, L07602, doi:10.1029/2011GL046757.

    • Search Google Scholar
    • Export Citation
  • Qu, T., , S. Gao, , and I. Fukumori, 2013: Formation of salinity maximum water and its contribution to the overturning circulation in the North Atlantic as revealed by a global general circulation model. J. Geophys. Res. Oceans, 118, 19821994, doi:10.1002/jgrc.20152.

    • Search Google Scholar
    • Export Citation
  • Reverdin, G., , E. Kestenare, , C. Frankignoul, , and T. Delcroix, 2007: Surface salinity in the Atlantic Ocean (30°S–50°N). Prog. Oceanogr., 73, 311340, doi:10.1016/j.pocean.2006.11.004.

    • Search Google Scholar
    • Export Citation
  • Ruddick, B., 1983: A practical indicator of stability of the water column to double-diffusive activity. Deep-Sea Res., 30A, 11051107, doi:10.1016/0198-0149(83)90063-8.

    • Search Google Scholar
    • Export Citation
  • Sasaki, Y. N., , N. Schneider, , N. Maximenko, , and K. Lebedev, 2010: Observational evidence for propagation of decadal spiciness anomalies in the North Pacific. Geophys. Res. Lett., 37, L07708, doi:10.1029/2010GL042716.

    • 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, doi:10.1175/1520-0485(1981)011<1015:FOTTSR>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Schmitt, R. W., 1999: Spice and the demon. Science, 283, 498499, doi:10.1126/science.283.5401.498.

  • Schneider, N., 2004: The response of tropical climate to the equatorial emergence of spiciness anomalies. J. Climate, 17, 10831095, doi:10.1175/1520-0442(2004)017<1083:TROTCT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Schneider, N., , A. J. Miller, , M. A. Alexander, , and C. Deser, 1999: Subduction of decadal North Pacific temperature anomalies: Observations and dynamics. J. Phys. Oceanogr., 29, 10561070, doi:10.1175/1520-0485(1999)029<1056:SODNPT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Schott, F., and et al. , 2003: The zonal currents and transports at 35°W in the tropical Atlantic. Geophys. Res. Lett., 30, 1349, doi:10.1029/2002GL016849.

    • Search Google Scholar
    • Export Citation
  • Stommel, H., 1979: Determination of watermass properties of water pumped down from the Ekman layer to the geostrophic flow below. Proc. Natl. Acad. Sci. USA, 76, 30513055, doi:10.1073/pnas.76.7.3051.

    • Search Google Scholar
    • Export Citation
  • Wallace, J. M., , C. Smith, , and Q. Jiang, 1990: Spatial patterns of atmosphere–ocean interaction in the northern winter. J. Climate, 3, 990998, doi:10.1175/1520-0442(1990)003<0990:SPOAOI>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Wang, C., , S. Dong, , and E. Munoz, 2010: Seawater density variations in the North Atlantic and the Atlantic meridional overturning circulation. Climate Dyn., 34, 953968, doi:10.1007/s00382-009-0560-5.

    • Search Google Scholar
    • Export Citation
  • Williams, R. G., 1991: The role of the mixed layer in setting the potential vorticity of the main thermocline. J. Phys. Oceanogr., 21, 18031814, doi:10.1175/1520-0485(1991)021<1803:TROTML>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Woods, J. D., 1985: The physics of pycnocline ventilation. Coupled Ocean-Atmosphere Models, J. C. J. Nihoul, Ed., Elsevier, 543–590.

  • Worthington, L. V., 1976: On the North Atlantic Circulation. Johns Hopkins University Press, 110 pp.

  • Yeager, S. G., , and W. G. Large, 2004: Late-winter generation of spiciness on subducted isopycnals. J. Phys. Oceanogr., 34, 15281547, doi:10.1175/1520-0485(2004)034<1528:LGOSOS>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Yeager, S. G., , and W. G. Large, 2007: Observational evidence of winter spice injection. J. Phys. Oceanogr., 37, 28952919, doi:10.1175/2007JPO3629.1.

    • Search Google Scholar
    • Export Citation
  • Zhang, D., , M. J. McPhaden, , and W. E. Johns, 2003: Observational evidence for flow between the subtropical and tropical Atlantic: The Atlantic subtropical cells. J. Phys. Oceanogr., 33, 17831797, doi:10.1175/2408.1.

    • Search Google Scholar
    • Export Citation
  • Zhang, L., , and T. Qu, 2014: Low-frequency variability of South Pacific Tropical Water from Argo. Geophys. Res. Lett., 41, 24412446, doi:10.1002/2014GL059490.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    Annual-mean sea surface salinity (psu) in the subtropical North Atlantic from the World Ocean Atlas 2013 (WOA13). The white contours indicate winter (March) surface density (kg m−3), and the heavy black contour indicates the 37.0-psu isohaline.

  • View in gallery

    Annual-mean (a) wind stress (N m−2, black arrows, reference arrow provided) and its curl (10−7 N m−3, shading) from the NCEP reanalyses and (b) winter (March) mixed layer depth (m, shading) and surface density (kg m−3 white contours) from WOA13 in the subtropical North Atlantic. The heavy contour indicates the zero wind stress curl line in (a) and the 37.0-psu isohaline in (b).

  • View in gallery

    Number of Argo profiles in (a) each 2° × 2° box and (b) over the entire subtropical (5°–40°N) North Atlantic for the period from January 2002 to August 2015.

  • View in gallery

    (a) Annual subduction rate (m yr−1) and its components due to (b) vertical pumping and (c) lateral induction in the subtropical North Atlantic (5°–40°N) from Argo and NCEP reanalyses for the period 2002–14. The contours represent the 37.0-psu isohaline.

  • View in gallery

    Subduction rate (Sv) in the North Atlantic (5°–40°N) against winter sea surface (a) density (kg m−3) and (b) salinity (psu).

  • View in gallery

    (a) Probability of vertical salinity maximum and (b) potential density (kg m−3), (c) salinity (psu), and (d) potential temperature (°C) of this salinity maximum (>36.5 psu) in the density range 24.5–26.5 kg m−3 from each individual profile at observed levels. Regions with probability > 0.5 are shown in (b), (c), and (d).

  • View in gallery

    (a) Magnitude and (b) phase of horizontal density ratio at the vertical salinity maximum in the region where the probability of the vertical salinity maximum is higher than 0.5.

  • View in gallery

    Mean salinity along the 25.5 kg m−3 density surface in the subtropical North Atlantic in (a) March and (b) September. The black contour indicates the 37.0-psu isohaline, and the white line shows the horizontal salinity maximum.

  • View in gallery

    Annual subduction rate (Sv) and its components due to vertical pumping and lateral induction compared with the winter (January–March) NAO index obtained from the NOAA Climate Prediction Center for the period 2007–14.

  • View in gallery

    Time series of winter (March) (a) WCS, SSD, and MLD and (b) SST, SSS, and SSD averaged within the 37.0-psu isohaline at the sea surface. All time series are normalized by their std dev, which are 0.22 × 10−7 N m−3 for WCS, 0.11 kg m−3 for SSD, 19 m for MLD, 0.40°C for SST, and 0.04 psu for SSS, respectively. SST in (b) is shown with a reversed sign for easier comparison.

  • View in gallery

    (left) Wind stress (vector) and its curl (shading) anomalies compared with (right) SST anomalies in March (top) 2010 and (bottom) 2012. The units are °C for SST and 10−7 N m−3 for wind stress curl anomalies.

  • View in gallery

    (left) Winter (March) MLD anomalies and (right) vertical distribution of zonal geostrophic velocity anomalies at 50°W averaged over March–August (top) 2010 and (bottom) 2012. The 8-yr mean values have been subtracted before plotting. Positive MLD anomalies represent the deepening of the mixed layer, and positive velocity anomalies represent strengthening of the westward flow. On the left panels, the white dashed lines indicate sea surface density, and the black dashed lines show the geographical location of the section used for the right panels. Units are meter for MLD and 10−2 m s−1 for velocity.

  • View in gallery

    Horizontal distribution of SSS (psu) anomalies in March (a) 2010 and (b) 2012 and vertical distribution of salinity anomalies across the horizontal salinity maximum along the 25.5 kg m−3 density surface during (c) March 2010–February 2011 and (d) March 2012–February 2013. The location of the section in (c) and (d) is marked by the white solid lines in (a) and (b).

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 131 131 18
PDF Downloads 112 112 18

North Atlantic Subtropical Underwater and Its Year-to-Year Variability in Annual Subduction Rate during the Argo Period

View More View Less
  • 1 International Pacific Research Center, School of Ocean and Earth Science and Technology, University of Hawai‘i at Mānoa, Honolulu, Hawaii, and Joint Institute for Regional Earth System Science and Engineering, University of California, Los Angeles, Los Angeles, California
  • | 2 International Pacific Research Center, School of Ocean and Earth Science and Technology, University of Hawai‘i at Mānoa, Honolulu, Hawaii
© Get Permissions
Full access

Abstract

Subtropical underwater (STUW) and its year-to-year variability in annual subduction rate are investigated using recently available Argo data in the North Atlantic. For the period of observation (2002–14), the mean annual subduction rate of the STUW is 7.3 ± 1.2 Sv (1 Sv = 106 m3 s−1) within the density range between 25.0 and 26.0 kg m−3. Once subducted, the STUW spreads in the subtropical gyre as a vertical salinity maximum. In the mean, the spatial changes in temperature and salinity of the STUW tend to compensate each other, and the density of the water mass remains rather stable near 25.5 kg m−3 in the southwestern part of the subtropical gyre. The annual subduction rate of the STUW varies from year to year, and most of this variability is due to lateral induction, which in turn is directly linked to the variability of the winter mixed layer depth. Through modulation of surface buoyancy, wind anomalies associated with the North Atlantic Oscillation are primarily responsible for this variability. Sea surface salinity anomalies in the formation region of the STUW are conveyed into the thermocline, but their westward propagation cannot be detected by the present data.

School of Ocean and Earth Science and Technology Contribution Number 9607 and International Pacific Research Center Contribution Number IPRC-1183.

Corresponding author address: Dr. Tangdong Qu, Joint Institute for Regional Earth System Science and Engineering, University of California, Los Angeles, 607 Charles E. Young Drive East, Los Angeles, CA 90095. E-mail: tangdong@ucla.edu

Abstract

Subtropical underwater (STUW) and its year-to-year variability in annual subduction rate are investigated using recently available Argo data in the North Atlantic. For the period of observation (2002–14), the mean annual subduction rate of the STUW is 7.3 ± 1.2 Sv (1 Sv = 106 m3 s−1) within the density range between 25.0 and 26.0 kg m−3. Once subducted, the STUW spreads in the subtropical gyre as a vertical salinity maximum. In the mean, the spatial changes in temperature and salinity of the STUW tend to compensate each other, and the density of the water mass remains rather stable near 25.5 kg m−3 in the southwestern part of the subtropical gyre. The annual subduction rate of the STUW varies from year to year, and most of this variability is due to lateral induction, which in turn is directly linked to the variability of the winter mixed layer depth. Through modulation of surface buoyancy, wind anomalies associated with the North Atlantic Oscillation are primarily responsible for this variability. Sea surface salinity anomalies in the formation region of the STUW are conveyed into the thermocline, but their westward propagation cannot be detected by the present data.

School of Ocean and Earth Science and Technology Contribution Number 9607 and International Pacific Research Center Contribution Number IPRC-1183.

Corresponding author address: Dr. Tangdong Qu, Joint Institute for Regional Earth System Science and Engineering, University of California, Los Angeles, 607 Charles E. Young Drive East, Los Angeles, CA 90095. E-mail: tangdong@ucla.edu

1. Introduction

The world’s highest (>37.0 psu) sea surface salinity (SSS) in an open ocean is observed in the subtropical North Atlantic (Fig. 1). This high-salinity water experiences variability on various time scales, including a linear trend over the past decades in its salinity (e.g., Boyer et al. 2005; Reverdin et al. 2007; Gordon and Giulivi 2008; Wang et al. 2010; Durack and Wijffels 2010; Qu et al. 2011; Bingham et al. 2012), which is believed to play a role in climate variability through its involvement in global thermohaline circulation (e.g., Häkkinen 1999; Zhang et al. 2003; Schott et al. 2003; Mignot and Frankignoul 2004; de Boer et al. 2008; Qu et al. 2013). The linear trend of SSS in the subtropical North Atlantic has been linked to the recent amplification of the hydrological cycle (e.g., Held and Soden 2006; IPCC 2007) and the global warming effect on Earth’s freshwater cycle (e.g., Durack and Wijffels 2010).

Fig. 1.
Fig. 1.

Annual-mean sea surface salinity (psu) in the subtropical North Atlantic from the World Ocean Atlas 2013 (WOA13). The white contours indicate winter (March) surface density (kg m−3), and the heavy black contour indicates the 37.0-psu isohaline.

Citation: Journal of Physical Oceanography 46, 6; 10.1175/JPO-D-15-0246.1

On average, the North Atlantic SSS maximum lies in the central subtropical gyre (Fig. 1), where vertical Ekman pumping is predominantly downward (Fig, 2a). Thus, some portion of this high-salinity water enters the thermocline by Ekman pumping and forms a vertical salinity maximum, the subtropical underwater (STUW), along its southwestward advective path by the gyre circulation (Worthington 1976). The formation of the STUW is further enhanced by lateral induction, reflecting the influence of horizontal advection across a sloping mixed layer (e.g., Stommel 1979; Woods 1985; Williams 1991; Qiu and Huang 1995). The mixed layer in the subtropical North Atlantic is generally deep (Fig. 2b), with a maximum depth exceeding 100 m in winter. As in many parts of the global ocean, the mixed layer depth (MLD) exhibits a large annual excursion, approaching its seasonal maximum in winter and seasonal minimum in summer. The rapid shoaling of the mixed layer in late spring allows winter mixed layer water to subduct into the thermocline.

Fig. 2.
Fig. 2.

Annual-mean (a) wind stress (N m−2, black arrows, reference arrow provided) and its curl (10−7 N m−3, shading) from the NCEP reanalyses and (b) winter (March) mixed layer depth (m, shading) and surface density (kg m−3 white contours) from WOA13 in the subtropical North Atlantic. The heavy contour indicates the zero wind stress curl line in (a) and the 37.0-psu isohaline in (b).

Citation: Journal of Physical Oceanography 46, 6; 10.1175/JPO-D-15-0246.1

Based on sporadic hydrographic and tracer data, earlier studies suggested that the annual subduction rate of the STUW ranges from 2 to 5 Sv in the North Atlantic (e.g., Qiu and Huang 1995; O’Connor et al. 2005), while results from ocean general circulation models suggest a slightly larger value (~6 Sv) within the density range of 25–26 kg m−3 (Qu et al. 2013). Once subducted, the STUW spreads as a vertical salinity maximum in the subtropical gyre, through which variability in the SSS maximum can be conveyed into the thermocline, contributing directly to the formation of the North Atlantic Deep Water and consequently the thermohaline circulation of the global ocean (e.g., Blanke et al. 2002; Wang et al. 2010; Burkholder and Lozier 2011; Qu et al. 2013).

Water properties of the North Atlantic STUW have been investigated by previous studies (e.g., Worthington 1976; O’Connor et al. 2005). Based on climatological temperature and salinity data, O’Connor et al. (2005) noted that the North Atlantic STUW lies around 26.0 kg m−3, with mean potential temperature and salinity of about 20.4°C and 36.7 psu, respectively. In the central subtropical gyre, near 20°N and 35°W, the salinity of the STUW exceeds 37.0 psu, which is the highest in an open ocean. The average age of the STUW as estimated from the tracer data is 1–5 yr, and after this period the water gradually loses its characteristics as a result of mixing (e.g., O’Connor et al. 2005; Laurian et al. 2009).

Though previous studies have significantly advanced our knowledge of the STUW in the North Atlantic, the water’s large-scale structure and year-to-year variability still remain poorly understood. The present study uses the unprecedented number of temperature and salinity profiles collected by Argo floats deployed in the past decade to examine the formation, circulation, and variability of the North Atlantic STUW. Combining the decade-long Argo data with surface wind and flux products, processes that govern the North Atlantic STUW variability are discussed.

The remainder of the paper offers a brief description of the data and method of analysis (section 2), followed by estimates of the annual subduction rate of the STUW (section 3). Variability of the subduction rate and its associated atmospheric forcing are then discussed (section 4), and the results are summarized (section 5).

2. Data description

To trace the North Atlantic STUW either as a vertical salinity maximum or along density surfaces, we use individual profiles, as well as a gridded product of Argo observations created by the Asian Pacific Data Research Center (APDRC) of the International Pacific Research Center (IPRC), University of Hawaii (called the APDRC product hereinafter). Individual profiles, which typically have a better vertical resolution than the gridded data, are used to detect the vertical salinity maximum associated with the STUW and enable a detailed description of the water’s characteristics. The APDRC product provides the three-dimensional context of the gyre circulation.

a. Individual profiles

Concurrent temperature–salinity profiles from Argo at observed levels are used for this study. The Argo floats record temperature and salinity from a typical top level around 5 to about 2000 m, with a vertical resolution of 1–5 m in the upper ocean and somewhat coarser at depth (25 m as the standard). We first remove the profiles that were flagged as “bad” or as not passing the monthly, seasonal, and annual standard deviation checks and then remove the profiles that extended shallower than 100 m, simultaneously eliminating profiles over the continental shelf. To correct or remove erroneous records in both coordinates and measured values, extensive hand editing of the data is then conducted. After these steps, 77 666 “reliable” temperature–salinity profiles remain in the North Atlantic between 5° and 40°N for the period from January 2002 to August 2015 (Fig. 3a). Over the region studied, the number of profiles is relatively low (~2000) in the early 2000s and increases to more than 10 000 month−1 in recent years (Fig. 3b).

Fig. 3.
Fig. 3.

Number of Argo profiles in (a) each 2° × 2° box and (b) over the entire subtropical (5°–40°N) North Atlantic for the period from January 2002 to August 2015.

Citation: Journal of Physical Oceanography 46, 6; 10.1175/JPO-D-15-0246.1

Potential temperature, salinity, and potential density at the vertical salinity maximum (>37.0 psu) are computed from each individual profile. Regardless of the date and time of observation, properties from individual profiles are then averaged in a 2° × 2° grid. To eliminate the bias in density of sampling, a variable scan radius is used so that each average is based on at least five samples (e.g., Qu et al. 1999). The gridded properties are smoothed using a Gaussian filter, with an e-folding scale of 200 km. Standard deviations are also calculated and used to edit the mean properties. Values from an individual Argo profile that deviate from the grid mean by 3 times the standard deviation, or more, are excluded from further analysis, and the grid mean and standard deviation are recalculated. The typical standard errors, or standard deviations divided by the square root of the number of samplings, of potential temperature, salinity, and potential density at the vertical salinity maximum, defined as the regional averages over the subtropical North Atlantic, are 0.15°C, 0.02 psu, and 0.04 kg m−3, respectively, based on ensembles of more than five samples in each 2° × 2° grid bin. These values will be used to measure the uncertainties of the mean property fields (see section 3).

b. APDRC product

The near–real time, monthly averaged temperature–salinity or APDRC data product, prepared by APDRC/IPRC of the University of Hawaii, covers the global ocean in a 1° × 1° grid on 26 standard levels in the upper 2000 m and spans the period from January 2005 to present. The standard levels used for the present study are the same as those for the World Ocean Atlas (e.g., Levitus 1982). In preparing this dataset, a variational analysis technique was used to interpolate temperature and salinity onto a three-dimensional spatial grid. For more details about this technique and this dataset, see the documentation presented online (at http://apdrc.soest.hawaii.edu/projects/Argo/index.php). The gridded temperature and salinity are then converted into dynamic height by assuming a reference level, from which geostrophic velocity can be further derived.

3. Mean characteristics of the STUW

a. Annual subduction rate

The annual subduction rate SRann, or volume flux of mixed layer water entering the thermocline per unit horizontal area, can be estimated by tracing water particles released at the base of the winter mixed layer for 1 yr (e.g., Qiu and Huang 1995); that is,
e1
where hm is the winter MLD, wmb is the vertical velocity of water parcel at the base of mixed layer resulting from Ekman pumping and geostrophic convergence, and t1 and t2 are the end of the first and second winter, respectively. The term T represents the time period of 1 yr. The first term on the right-hand side of Eq. (1) is the contribution of vertical pumping. The second term is the contribution of lateral induction, reflecting the influence of horizontal advection against the shoaling mixed layer [Qiu and Huang (1995) and references therein].

The mean seasonal cycles of the MLD and geostrophic velocity used for the present analysis are derived from the APDRC product. Here, the MLD is calculated as the depth where σθ is equal to the sea surface σθ plus an increment of 0.1 kg m−3. Different criteria can be used to define the MLD, provided the question of sensitivity is born in mind (e.g., Karstensen and Quadfasel 2002). Recent studies have shown that changes in the σθ increment between 0.075 and 0.2 kg m−3 can give up to 20% changes in MLD, resulting in an uncertainty of about 6% in the STUW subduction rate (Lu et al. 2016, manuscript submitted to J. Geophys. Res.). So, to a large extent, the STUW subduction rate is insensitive to the selection of MLD criteria. The reference level for geostrophic velocity is chosen at 2000 m, which is the deepest possible common depth of the data. Combining these hydrographic data with the wind product from the NCEP reanalyses (Kalnay et al. 1996), we are able to estimate the annual subduction rate in the subtropical North Atlantic (Fig. 4a), which shows essentially the same pattern as earlier studies (e.g., Qiu and Huang 1995; Qu et al. 2013), though the newly available Argo data capture subduction in a detail unseen previously. Two bands of high subduction rate are shown in the subtropical North Atlantic. The southern one is associated with the STUW subduction, where surface salinity is typically higher than 37.0 psu.

Fig. 4.
Fig. 4.

(a) Annual subduction rate (m yr−1) and its components due to (b) vertical pumping and (c) lateral induction in the subtropical North Atlantic (5°–40°N) from Argo and NCEP reanalyses for the period 2002–14. The contours represent the 37.0-psu isohaline.

Citation: Journal of Physical Oceanography 46, 6; 10.1175/JPO-D-15-0246.1

The total subduction rate in the subtropical North Atlantic between 5° and 40°N is 33.1 Sv (1 Sv = 106 m3 s−1). Over this region, the annual subduction rate due to vertical pumping is typically larger than 40 m yr−1 (Fig. 4b), while its counterpart due to lateral induction is narrowly confined to the central part of the subtropical gyre (Fig. 4c), where its magnitude reaches 20–30 m yr−1. Averaged over the SSS maximum region enclosed by the 37.0-psu isohaline, approximately 68% of the total subduction rate is forced by vertical pumping, which dominates over the contribution (32%) from lateral induction, in agreement with earlier studies based on tracer data (e.g., O’Connor et al. 2005) and numerical models (e.g., Qu et al. 2013).

When plotted against winter (March) surface density, one can isolate the subduction of the STUW that is supposed to occur between 25.0 and 26.0 kg m−3 (Fig. 5a). Earlier studies have shown that the subduction rate of the STUW ranges from 2 to 6 Sv (e.g., Qiu and Huang 1995; O’Connor et al. 2005; Qu et al. 2013). Our result (7.3 Sv) favors a slightly higher value than these earlier estimates. If a density range of 25.6–26.3 kg m−3 is used (O’Connor et al. 2005), the subduction rate of the STUW reaches 8.6 Sv. The subduction rate of the STUW is also estimated based on its high-salinity (>37.0 psu) signature (Fig. 5b), and our estimate (5.2 Sv) from the Argo data is smaller by a factor of about 2 than that (11.3 Sv) reported by Blanke et al. (2002), who used a broader formation area enclosed by the 36.6-psu isohaline.

Fig. 5.
Fig. 5.

Subduction rate (Sv) in the North Atlantic (5°–40°N) against winter sea surface (a) density (kg m−3) and (b) salinity (psu).

Citation: Journal of Physical Oceanography 46, 6; 10.1175/JPO-D-15-0246.1

There are uncertainties in estimating the subduction rate using Eq. (1). Most of them result from the selection of reference level and MLD criterion. By conducting a number of sensitivity tests, Lu et al. (2016, manuscript submitted to J. Geophys. Res.) noted that these uncertainties account for less than 8.5% of the STUW subduction rate in the South Pacific, which is significantly smaller than its year-to-year variability. In the North Atlantic, the standard deviation of the STUW subduction rate derived from the 8-yr time series reaches 1.2 and 2.5 Sv within the 25.0–26.0 kg m−3 and 25.6–26.3 kg m−3 density range, respectively. If a similar rate (<8.5%) applies, these standard deviations exceed the uncertainties noted above by a factor of 2.0–3.5, suggesting that the subduction rate of the STUW and its year-to-year variability presented in this study are robust.

b. Mean property distribution

Once subducted, the STUW spreads as part of the gyre circulation. In the present study, the North Atlantic STUW is first traced as a vertical salinity maximum (>36.5 psu) in the density range between 24.5 and 26.5 kg m−3, using individual Argo profiles. In the presence of mixing, using the salinity maximum as an indicator may better preserve the STUW properties than using isopycnal surfaces (e.g., Qu et al. 1999). Where multiple maxima occur, the one of highest salinity is selected. The probability of the salinity maximum within the density range cited above, defined as the ratio of the number of profiles that contain at least one salinity maximum to the total number of profiles in each grid bin (Qu et al. 1999), shows the spreading of the STUW (Fig. 6a), with values higher than 0.5 covering a large part of the subtropical gyre. The highest (>0.9) probability is seen to extend westward from the formation region of the STUW. Upon reaching the western boundary, most of the STUW turns northward, while its equatorward flow is weak, consistent with the previous studies (e.g., Zhang et al. 2003; Schott et al. 2003; Qu et al. 2013). The high (>0.9) probability belt stretching northeastward off the east coast of America (Fig. 6a) and farther eastward in the northern subtropical gyre suggests recirculation of the STUW (e.g., Qu et al. 2013).

Fig. 6.
Fig. 6.

(a) Probability of vertical salinity maximum and (b) potential density (kg m−3), (c) salinity (psu), and (d) potential temperature (°C) of this salinity maximum (>36.5 psu) in the density range 24.5–26.5 kg m−3 from each individual profile at observed levels. Regions with probability > 0.5 are shown in (b), (c), and (d).

Citation: Journal of Physical Oceanography 46, 6; 10.1175/JPO-D-15-0246.1

The density of the STUW varies between 25.0 and 26.0 kg m−3 with an average of 25.5 kg m−3 over the domain enclosed by the probability higher than 0.5 (Fig. 6b). The water is slightly denser around its formation region and becomes lighter once subducted. Upon approaching the western boundary, a majority of this water turns northwestward and then northeastward. As it recirculates in the northern subtropical gyre, its density gradually increases. This result is consistent with earlier studies (e.g., Qu et al. 2013), suggesting that the North Atlantic STUW is variable in density as a result of mixing.

Considerable spatial changes are observed in salinity and potential temperature of the vertical salinity maximum. Salinity of the STUW is highest (>37.4 psu) in the central subtropical gyre, where the North Atlantic STUW is formed, and drops gradually toward the southwest (Fig. 6c). Along the western boundary, a tongue of high (>36.8 psu) salinity extends northwestward. Some of this high-salinity water turns eastward and recirculates in the subtropical gyre, while the rest continues northeastward, contributing directly to the formation of North Atlantic Deep Water (e.g., Qu et al. 2013). Potential temperature is high (>23°C) in the formation region of the STUW, decreases southwestward in the southern flank of the subtropical gyre, and falls below 21°C near the western boundary (Fig. 6d). It is worth noting that, despite these spatial changes in potential temperature and salinity, the density of the STUW remains stable in the southwestern part of the subtropical gyre, south of 25°N and west of 30°W (Fig. 6b). The vertical salinity maximum associated with the STUW lies in a narrow density range between 25.3 and 25.4 kg m−3. Farther to the north, the density at the vertical salinity maximum exceeds 26.0 kg m−3. Note that the typical standard errors of potential temperature, salinity, and potential density at the vertical salinity maximum (see section 2) are significantly smaller than the contour intervals chosen for Fig. 6. So, statistically, most of the STUW structures described above are robust.

To further demonstrate the density compensation of temperature and salinity in the STUW, we examine the complex horizontal density ratio RH (Ruddick 1983; Yeager and Large 2007; Kolodziejczyk et al. 2015), defined as
eq1
where α and β are the expansion coefficient of temperature and the contraction coefficient of salinity, and θx (Sx) and θy (Sy) are the gradient of potential temperature (salinity) of the vertical salinity maximum in the x and y direction, respectively. The magnitude and phase of RH denote the strength ratio and difference in direction, respectively, of the horizontal density gradients due to temperature and salinity. A phase close to zero indicates density-compensating temperature and salinity gradients.

Horizontal potential temperature and salinity gradients in the STUW are nearly aligned with direction difference less than 20° (Fig. 7). This result indicates that changes in potential temperature and salinity in the vertical salinity maximum tend to compensate each other (Fig. 7). Such compensation in the formation region of the STUW was previously reported by Kolodziejczyk et al. (2015), and our result suggests that the compensation remains effective after the water is subducted, consistent with the potential density distribution shown in Fig. 6b. So, to a large extent, the North Atlantic STUW can be traced as a passive tracer along isopycnal surfaces.

Fig. 7.
Fig. 7.

(a) Magnitude and (b) phase of horizontal density ratio at the vertical salinity maximum in the region where the probability of the vertical salinity maximum is higher than 0.5.

Citation: Journal of Physical Oceanography 46, 6; 10.1175/JPO-D-15-0246.1

Figure 8 shows the salinity distribution along the 25.5 kg m−3 isopycnal surface in winter (March) and summer (September). The spreading of the high-salinity STUW in both seasons can be identified as a “river of salt” (>37.0 psu) in the southern subtropical gyre. In winter (Fig. 8a), the 25.5 kg m−3 isopycnal surface outcrops, and the river of salt is directly ventilated to the surface mixed layer. In summer (Fig. 8b), when relatively warm and light water overlies, the river of salt (>37.0 psu) extends farther northward in the seasonal thermocline. Interestingly, salinity both along the 25.5 kg m−3 isopycnal surface (Fig. 8) and at the vertical salinity maximum (Fig. 6c) shows a northeast–southwest-oriented distribution, different from that at the sea surface (Fig. 1), reflecting strong influence of ocean processes (e.g., Qu et al. 2011).

Fig. 8.
Fig. 8.

Mean salinity along the 25.5 kg m−3 density surface in the subtropical North Atlantic in (a) March and (b) September. The black contour indicates the 37.0-psu isohaline, and the white line shows the horizontal salinity maximum.

Citation: Journal of Physical Oceanography 46, 6; 10.1175/JPO-D-15-0246.1

4. Year-to-year variability and its related forcing mechanism

a. Variability of subduction rate

Winter (March) MLD and monthly geostrophic velocity derived from the APDRC product are used to examine the year-to-year variability of the STUW annual subduction rate [Eq. (1)]. Since the data coverage at the beginning of the Argo program is relatively sparse (Fig. 3b), our discussion below focuses on the period 2007–14. Over this period of time, the pattern of the annual subduction rate during each individual year is similar to the multiyear average (Fig. 4), with high values in the central subtropical gyre corresponding to the formation of the STUW (figure not shown). However, the volume of the STUW subducted within the 37.0-psu isohaline varies greatly from year to year (Fig. 9), with a minimum of 3.1 Sv in 2010 and a maximum of 6.9 Sv in 2014. A relatively large (>6.0 Sv) subduction rate is also found in 2008 and 2009.

Fig. 9.
Fig. 9.

Annual subduction rate (Sv) and its components due to vertical pumping and lateral induction compared with the winter (January–March) NAO index obtained from the NOAA Climate Prediction Center for the period 2007–14.

Citation: Journal of Physical Oceanography 46, 6; 10.1175/JPO-D-15-0246.1

The variability of the subduction rate is primarily driven by lateral induction, even though vertical pumping is dominant for its mean value (section 3a). During the period 2007–14, vertical pumping remains nearly unchanged (Fig. 9), with its standard deviation (0.5 Sv) accounting for only about 14% of its mean value (3.5 Sv). The year-to-year variability of lateral induction is much more significant, and its standard deviation (1.4 Sv) is about 74% of its mean value (1.9 Sv). This variability clearly shows a North Atlantic Oscillation (NAO) signal (Barnston and Livezey 1987), with enhanced value in 2008/09 and 2011/12 and reduced value in 2010/11 corresponding well with the NAO index obtained from the NOAA Climate Prediction Center (Fig. 9).

b. Governing processes

As is well known, winter MLD is a key process modulating the subduction rate of the STUW through lateral induction [Eq. (1)]. Examination of the APDRC data product shows that the winter MLD in the formation region of the STUW experiences a large variability in the past decade (Fig. 10a). Averaged within the region of S > 37.0 psu at the sea surface, this variability is nearly in phase with the NAO index during the period 2007–14 (Fig. 9). It shoals during the negative phase of the NAO and deepens during the positive phase of the NAO. Good examples of this variability are the MLD minimum in winter 2010 and maximum in winter 2012, when the NAO reaches its weakest and strongest strength, respectively (Fig. 9).

Fig. 10.
Fig. 10.

Time series of winter (March) (a) WCS, SSD, and MLD and (b) SST, SSS, and SSD averaged within the 37.0-psu isohaline at the sea surface. All time series are normalized by their std dev, which are 0.22 × 10−7 N m−3 for WCS, 0.11 kg m−3 for SSD, 19 m for MLD, 0.40°C for SST, and 0.04 psu for SSS, respectively. SST in (b) is shown with a reversed sign for easier comparison.

Citation: Journal of Physical Oceanography 46, 6; 10.1175/JPO-D-15-0246.1

The deepening (shoaling) of the MLD represents a downward (upward) buoyancy flux or a conversion between turbulent kinetic energy and potential energy, driven by surface wind stirring and/or surface convection (e.g., Niiler 1975). On a large scale, it is also influenced by surface wind stress curl (WSC), with deeper MLD corresponding to negative WSC, and vice versa (e.g., Qu and Chen 2009; Liu and Huang 2012). Such a correspondence in the formation region of the South Pacific STUW was reported by Lu et al. (2016, manuscript submitted to J. Geophys. Res.), but it does not hold for the North Atlantic (Fig. 10a). In fact, averaged within the 37.0-psu isohaline at the sea surface, a slightly positive correlation is seen between the winter MLD and WSC in the subtropical North Atlantic, which is against the correspondence noted above. The implication of this result is that local Ekman pumping cannot be a major cause of the MLD variability in the region studied.

Also included in Fig. 10a is a time series of normalized sea surface density (SSD) averaged within the 37.0-psu isohaline in winter (March). During the period 2007–14, the normalized winter SSD and MLD time series are nearly identical, with higher SSD corresponding to deeper MLD and lower SSD corresponding to shallower MLD. This result suggests that buoyancy flux stemming from surface wind stirring and/or convection is primarily responsible for the winter MLD variability in the formation region of the STUW in the North Atlantic. Further inspection of the SST and SSS time series indicates that the SSD variability is mostly thermal driven (Fig. 10b). In other words, it is the SST variability that controls the winter MLD in the formation region of the STUW, which in turn is directly linked to the NAO. Here, we note that the impact of SSS variability on the MLD is weak but not negligible (Fig. 10b). For example, a noticeable SSS jump took place between 2010 and 2012. Averaged within the 37.0-psu isohaline at the sea surface, this SSS jump (~0.07 psu) accounts for 17% of the SSD increase (0.33 kg m−3), directly contributing to the deepening of MLD in the formation region of the STUW during the period of observation (Fig. 10b). We will return to this point later.

The joint occurrence of large-scale SST and wind anomalies in the northern winter has been reported by previous studies (e.g., Wallace et al. 1990; Deser and Timlin 1997). Our discussion below focuses on two winter events, one in 2010 and the other in 2012, corresponding to a negative and a positive phase of the NAO, respectively (Fig. 9a). In winter (March) 2010, when a negative phase of the NAO develops, southwesterly wind anomalies intensify over much of the subtropical North Atlantic (Fig. 11a), leading to a weakening of the trade winds. A weaker than normal heat loss associated with lower trade winds partially explains the positive SST anomalies in the southeastern part of the subtropical gyre (Fig. 11c). In contrast, enhanced cooling resulting from stronger trade winds in winter 2012 (Fig. 11b) contributes to the negative SST anomalies there (Fig. 11d). In the northwestern part of the subtropical gyre, the situation is generally reversed, where westerly winds are enhanced in winter 2010 and weakened in winter 2012, in response to the basin-scale wind anomalies associated with the NAO (Figs. 11a,b). The corresponding variability in evaporation and precipitation (figure not shown) also affects SSS (Fig. 10b), playing a role in regulating the surface salinity maximum in the North Atlantic (e.g., Mignot and Frankignoul 2004; Qu et al. 2011; D’Addezio and Bingham 2014). But, as noted above, its direct contribution to the SSD variability is relatively minor (Fig. 10b). Most of the SSD variability in the formation region of the STUW, lighter in 2010 and denser in 2012, is due to SST variability, which in turn is forced by wind anomalies associated with the NAO.

Fig. 11.
Fig. 11.

(left) Wind stress (vector) and its curl (shading) anomalies compared with (right) SST anomalies in March (top) 2010 and (bottom) 2012. The units are °C for SST and 10−7 N m−3 for wind stress curl anomalies.

Citation: Journal of Physical Oceanography 46, 6; 10.1175/JPO-D-15-0246.1

To further demonstrate how these surface processes influence the formation of the STUW, Figs. 12a and 12b show the MLD anomalies from its 8-yr mean value in the subtropical North Atlantic in winter 2010 and 2012. Two bands of large MLD variability are visible in the region. The northern one lies in the northwestern part of the subtropical gyre, and the southern one is located in the central part of the subtropical gyre. In winter 2010, MLD deepens in the northern band but shoals in the southern band by up to 70 m in magnitude (Fig. 12a). The situation is reversed in 2012 (Fig. 12b). This variability in MLD directly influences the subduction rate of the STUW through lateral induction [Eq. (1)]. In winter 2012, for example, as the positive phase of the NAO develops (Fig. 11b), deeper than normal MLD resulting from anomalously cold SST and dense SSD (Fig. 12b) allows mixed layer water to be easily swept into the thermocline by horizontal circulation, enhancing the subduction rate of the STUW. The opposite takes place in 2010, when the negative phase of NAO develops.

Fig. 12.
Fig. 12.

(left) Winter (March) MLD anomalies and (right) vertical distribution of zonal geostrophic velocity anomalies at 50°W averaged over March–August (top) 2010 and (bottom) 2012. The 8-yr mean values have been subtracted before plotting. Positive MLD anomalies represent the deepening of the mixed layer, and positive velocity anomalies represent strengthening of the westward flow. On the left panels, the white dashed lines indicate sea surface density, and the black dashed lines show the geographical location of the section used for the right panels. Units are meter for MLD and 10−2 m s−1 for velocity.

Citation: Journal of Physical Oceanography 46, 6; 10.1175/JPO-D-15-0246.1

Another important process that influences the subduction rate of the STUW is the variability of gyre circulation. Using the APDRC monthly product and assuming a reference level at 2000 db, we calculate the geostrophic velocity of the North Equatorial Current (NEC) between 5° and 25°N at 50°W as a major component of the subtropical gyre. Anomalies of geostrophic velocity from its 8-yr mean values during the two winter events are shown in Figs. 12c and 12d. Consistent with large-scale variability in wind stress curl (Fig. 11), the westward-flowing NEC is considerably weaker in 2010 and stronger in 2012. This variability in geostrophic velocity contributes to the variability of subduction rate, but it only explains about 3% of the difference between 2012 and 2010, suggesting that the MLD variability is the primary process modulating the subduction rate of the STUW in the North Atlantic.

c. Subduction of salinity anomalies

The question that arises immediately from the above analyses is if the variability in the SSS maximum can be conveyed into the thermocline. Similar questions have been raised by previous studies for other subtropical oceans (e.g., Sasaki et al. 2010; Chen et al. 2010; Zhang and Qu 2014). To address this question, Figs. 13a and 13b show the SSS anomalies in winter 2010 and 2012 (Figs. 13a,b). As already discussed above (Fig. 10b), SSS in the formation region of the STUW is generally lower in 2010 and higher in 2012. Averaged within the 37.0-psu isohaline at the sea surface, this salinity difference between the two winters exceeds 0.07 psu (Fig. 10b). The higher than usual (~37.8 psu) SSS in the formation region of the STUW in fall 2012 was recently reported by the Salinity Processes in the Upper Ocean Regional Study (SPURS; e.g., Bingham et al. 2014; Hormann et al. 2015; Kolodziejczyk et al. 2015). In a quantitative sense, our estimate (37.6 psu) is slightly lower than the SPURS measurement, presumably due to the averaging and smoothing of the Argo data, but the year-to-year variability of SSS from the Argo data is still markedly evident (Figs. 13a,b).

Fig. 13.
Fig. 13.

Horizontal distribution of SSS (psu) anomalies in March (a) 2010 and (b) 2012 and vertical distribution of salinity anomalies across the horizontal salinity maximum along the 25.5 kg m−3 density surface during (c) March 2010–February 2011 and (d) March 2012–February 2013. The location of the section in (c) and (d) is marked by the white solid lines in (a) and (b).

Citation: Journal of Physical Oceanography 46, 6; 10.1175/JPO-D-15-0246.1

To explore the horizontal extension of SSS anomalies after they are subducted, Figs. 13c and 13d show the vertical distribution of salinity anomalies along the horizontal salinity maximum on the 25.5 kg m−3 density surface during March 2010–February 2011 and March 2012–February 2013, respectively. In both events, salinity anomalies extend westward in the density range between 25.0 and 26.0 kg m−3. They are negative in 2010 and positive in 2012, consistent with the SSS variability shown in Figs. 13a and 13b. Some of these salinity anomalies are believed to be density compensated by temperature anomalies (e.g., Kolodziejczyk et al. 2015), often referred to as spiciness anomalies (e.g., Schmitt 1999), and advected by mean circulation in the subtropical gyre (e.g., Lazar et al. 2001; Yeager and Large 2004; Luo et al. 2005), eventually affecting the thermocline structure in the tropics (Gu and Philander 1997; Schneider et al. 1999; Schneider 2004; Kolodziejczyk and Gaillard 2012) and at high latitudes (e.g., Laurian et al. 2006; Qu et al. 2013).

However, with the available Argo data, we are unable to detect the westward propagation of these salinity anomalies. After a careful examination of available Argo profiles in each individual month, we find that salinity anomalies injected in the formation region of the STUW are confined to a limited longitude band mostly east of 50°W. The causes of the erosion farther westward are not clear. They could result from the sparseness of data coverage in the western part of the basin (Fig. 2a) or from enhanced ocean mixing associated with the western boundary current. The latter may be linked to the anomalous geostrophic current acting against mean salinity gradients or anomalous geostrophic advection (e.g., Kilpatrick et al. 2011). It may also be linked to salt fingering, a double-diffusive type of convective mixing driven by the fact that heated water diffuses more readily than salty water (e.g., Schmitt 1981; Busecke et al. 2014). Both processes deserve further investigation.

5. Summary

Based on individual temperature–salinity profiles and a gridded data product from Argo, we are able to show the large-scale structure of the North Atlantic STUW in a detail that is unseen previously. The decade-long time series of the Argo data also allows us to reveal a year-to-year variability of the North Atlantic STUW subduction rate, which is consistent with the upper-ocean conditions reported by previous studies (e.g., Bingham et al. 2014; Kolodziejczyk and Gaillard 2012; Kolodziejczyk et al. 2015) but has never been explicitly explored. A brief summary of the results is given below.

We first trace the North Atlantic STUW as a vertical salinity maximum in the upper thermocline, using available Argo profiles at observed levels. This criterion is new compared with those using isopycnal surfaces (e.g., O’Connor et al. 2005). In the presence of mixing, using the salinity maximum as a criterion may better preserve the STUW properties (e.g., Qu et al. 1999), thus allowing for a more accurate description of this water mass in the subtropical North Atlantic. Our analysis shows that despite its large horizontal temperature and salinity gradients, the STUW density remains rather stable in the southern part of the subtropical gyre. On the annual average, spatial changes in temperature and salinity of the STUW tend to compensate each other, leading to a horizontal density ratio well between 0.5 and 2.0 after the water is subducted. To a large extent, the North Atlantic STUW can be nicely traced along density surfaces near 25.5 kg m−3.

The annual subduction rate of the STUW within the density range between 25.0 and 26.0 kg m−3 is 7.3 Sv, with a standard deviation of 1.2 Sv against its mean value over the Argo period. This estimate compares reasonably well with earlier studies based on climatological data and model outputs (e.g., Qiu and Huang 1995; O’Connor et al. 2005; Qu et al. 2013). The decade-long time series of Argo data further allows us to examine the year-to-year variability of the STUW in annual subduction rate and its related causal mechanism in the North Atlantic. It is shown that most of the variability in the STUW subduction rate is due to lateral induction, despite the dominant role of vertical pumping in determining its mean value. As is well known, variability in winter MLD is the primary process influencing the lateral induction of the STUW, which in turn is directly linked to the buoyancy fluxes at the sea surface.

Variability of the STUW subduction rate shows a strong NAO signal, suggesting that large-scale atmospheric circulation is a major cause of this variability. Two winter events, representing the negative and positive phase of the NAO, respectively, are chosen to show how wind anomalies associated with the NAO generate variability of the STUW subduction rate in the North Atlantic. Note that the influence of wind anomalies associated with the NAO on sea surface properties of the North Atlantic has been reported by previous studies (e.g., Wallace et al. 1990; Deser and Timlin 1997). Recently, using Argo data, Kolodziejczyk et al. (2015) demonstrate that this influence extends to the thermocline and directly contributes to the winter MLD variability. The present analysis confirms these earlier results and shows further evidence for the NAO influence on the STUW subduction rate in the North Atlantic.

Our result also shows that the SSS anomalies in the formation region of the STUW can be conveyed into the thermocline, through which the NAO can have an imprint in the STUW. Once subducted, these salinity anomalies extend westward in the subtropical gyre, but they are confined to a very limited longitude band, mostly to the east of about 50°W. The sparse data coverage in the region does not allow us to detect the propagation signal west of this longitude. Among other things, the erosion of this propagation signal can be partially attributed to anomalous geostrophic advection and salt fingering. We will leave this topic for future studies.

Finally, we note that the 2010 winter event, characterized by a historically low NAO index, coincided with the reversal of the strong 2009/10 El Niño event. This suggests a probable winter teleconnection between the Pacific El Niño event and SST anomalies of the same sign in the tropical North Atlantic (e.g., Giannini et al. 2001). The opposite occurred during the 2012 winter event, when the higher NAO index and the reversal of the 2011/12 La Niña event concurred (e.g., Kolodziejczyk et al. 2015). The details require further investigation.

Acknowledgments

This study was supported by NASA as part of the Aquarius Science Team investigation through Grants NNX16AE10G and NNXl4AH38G and by JAMSTEC through its sponsorship of the JAMSTEC-IPRC Collaborative Research project. The authors are grateful to I. Fukumori, J. Busecke, and F. Bingham for useful discussion on the topic and to two anonymous reviewers for thoughtful comments and suggestions on an earlier version of the manuscript. (The individual Argo profiles were obtained from http://www.nodc.noaa.gov/OC5/WOD13, the gridded Argo products were obtained from http://apdrc.soest.hawaii.edu/, and the NAO index was obtained from http://www.cpc.ncep.noaa.gov/data/teledoc/nao.shtml.)

REFERENCES

  • Barnston, A. G., , and R. E. Livezey, 1987: Classification, seasonality and persistence of low-frequency atmospheric circulation patterns. Mon. Wea. Rev., 115, 10831126, doi:10.1175/1520-0493(1987)115<1083:CSAPOL>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Bingham, F. M., , G. R. Foltz, , and M. J. McPhaden, 2012: Characteristics of the seasonal cycle of surface layer salinity in the global ocean. Ocean Sci., 8, 915929, doi:10.5194/os-8-915-2012.

    • Search Google Scholar
    • Export Citation
  • Bingham, F. M., , J. Busecke, , A. L. Gordon, , C. F. Giulivi, , and Z. Li, 2014: The North Atlantic subtropical surface salinity maximum as observed by Aquarius. J. Geophys. Res. Oceans, 119, 77417755, doi:10.1002/2014JC009825.

    • Search Google Scholar
    • Export Citation
  • Blanke, B., , M. Arhan, , A. Lazar, , J. Mignot, , and G. Prévost, 2002: A Lagrangian numerical investigation of the origins and fates of the salinity maximum water in the Atlantic. J. Geophys. Res., 107, 3163, doi:10.1029/2002JC001318.

    • Search Google Scholar
    • Export Citation
  • Boyer, T. P., , S. Levitus, , J. Antonov, , R. Locarnini, , and H. Garcia, 2005: Linear trends in salinity for the World Ocean, 1955–1998. Geophys. Res. Lett., 32, L01604, doi:10.1029/2004GL021791.

    • Search Google Scholar
    • Export Citation
  • Burkholder, K. C., , and M. S. Lozier, 2011: Subtropical to subpolar pathways in the North Atlantic: Deductions from Lagrangian trajectories. J. Geophys. Res., 116, C07017, doi:10.1029/2010JC006697.

    • Search Google Scholar
    • Export Citation
  • Busecke, J., , Z. Li, , F. M. Bingham, , A. L. Gordon, , and J. Font, 2014: Mesoscale turbulence within the subtropical North Atlantic surface layer. J. Geophys. Res. Oceans, 119, 41244140, doi:10.1002/2013JC009715.

    • Search Google Scholar
    • Export Citation
  • Chen, J., , T. Qu, , Y. Sasaki, , and N. Schneider, 2010: Anti-correlated variability in subduction rate of the western and eastern North Pacific Oceans as seen from an eddy-resolving ocean GCM. Geophys. Res. Lett., 37, L23608, doi:10.1029/2010GL045239.

    • Search Google Scholar
    • Export Citation
  • D’Addezio, J. M., , and F. M. Bingham, 2014: A subtropical North Atlantic regional atmospheric moisture budget. J. Geophys. Res. Oceans, 119, 87318748, doi:10.1002/2014JC010300.

    • Search Google Scholar
    • Export Citation
  • de Boer, A. M., , J. R. Toggweiler, , and D. M. Sigman, 2008: Atlantic dominance of the meridional overturning circulation. J. Phys. Oceanogr., 38, 435450, doi:10.1175/2007JPO3731.1.

    • Search Google Scholar
    • Export Citation
  • Deser, C., , and M. S. Timlin, 1997: Atmosphere–ocean interaction on weekly timescales in the North Atlantic and Pacific. J. Climate, 10, 393408, doi:10.1175/1520-0442(1997)010<0393:AOIOWT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Durack, P., , and S. Wijffels, 2010: Fifty-year trends in global ocean salinities and their relationship to broad-scale warming. J. Climate, 23, 43424362, doi:10.1175/2010JCLI3377.1.

    • Search Google Scholar
    • Export Citation
  • Giannini, A., , J. C. H. Chiang, , M. A. Cane, , Y. Kushnir, , and R. Seager, 2001: The ENSO teleconnection to the tropical Atlantic Ocean: Contributions of the remote and local SSTs to rainfall variability in the tropical Americas. J. Climate, 14, 45304544, doi:10.1175/1520-0442(2001)014<4530:TETTTT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Gordon, A., , and C. F. Giulivi, 2008: Sea surface salinity trends. Oceanography, 21, 2029, doi:10.5670/oceanog.2008.64.

  • Gu, D., , and S. G. H. Philander, 1997: Interdecadal climate fluctuations that depend on exchanges between the tropics and extratropics. Science, 275, 805807, doi:10.1126/science.275.5301.805.

    • Search Google Scholar
    • Export Citation
  • Häkkinen, S., 1999: Variability of the simulated meridional heat transport in the North Atlantic for the period 1951–1993. J. Geophys. Res., 104, 10 99111 008, doi:10.1029/1999JC900034.

    • Search Google Scholar
    • Export Citation
  • Held, I. M., , and B. J. Soden, 2006: Robust responses of the hydrological cycle to global warming. J. Climate, 19, 56865699, doi:10.1175/JCLI3990.1.

    • Search Google Scholar
    • Export Citation
  • Hormann, V., , L. R. Centurioni, , and G. Reverdin, 2015: Evaluation of drifter salinities in the subtropical North Atlantic. J. Atmos. Oceanic Technol., 32, 185192, doi:10.1175/JTECH-D-14-00179.1.

    • Search Google Scholar
    • Export Citation
  • IPCC, 2007: Climate Change 2007: The Physical Science Basis. Cambridge University Press, 996 pp.

  • Kalnay, E., and et al. , 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77, 437371, doi:10.1175/1520-0477(1996)077<0437:TNYRP>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Karstensen, J., , and D. Quadfasel, 2002: Water subducted into the Indian Ocean subtropical gyre. Deep-Sea Res. II, 49, 14411457, doi:10.1016/S0967-0645(01)00160-6.

    • Search Google Scholar
    • Export Citation
  • Kilpatrick, T., , N. Schneider, , and E. Di Lorenzo, 2011: Generation of low-frequency spiciness variability in the thermocline. J. Phys. Oceanogr., 41, 365377, doi:10.1175/2010JPO4443.1.

    • Search Google Scholar
    • Export Citation
  • Kolodziejczyk, N., , and F. Gaillard, 2012: Observation of spiciness interannual variability in the Pacific pycnocline. J. Geophys. Res., 117, C12018, doi:10.1029/2012JC008365.

    • Search Google Scholar
    • Export Citation
  • Kolodziejczyk, N., , G. Reverdin, , and A. Lazar, 2015: Interannual variability of the mixed layer winter convection and spice injection in the eastern subtropical North Atlantic. J. Phys. Oceanogr., 45, 504525, doi:10.1175/JPO-D-14-0042.1.

    • Search Google Scholar
    • Export Citation
  • Laurian, A., , A. Lazar, , G. Reverdin, , K. Rodgers, , and P. Terray, 2006: Poleward propagation of spiciness anomalies in the North Atlantic Ocean. Geophys. Res. Lett., 33, L13603, doi:10.1029/2006GL026155.

    • Search Google Scholar
    • Export Citation
  • Laurian, A., , A. Lazar, , and G. Reverdin, 2009: Generation mechanism of spiciness anomalies: An OGCM analysis in the North Atlantic Subtropical Gyre. J. Phys. Oceanogr., 39, 10031018, doi:10.1175/2008JPO3896.1.

    • Search Google Scholar
    • Export Citation
  • Lazar, A., , R. Murtugudde, , and A. J. Busalacchi, 2001: A model study of temperature anomaly propagation from subtropics to tropics within the South Atlantic thermocline. Geophys. Res. Lett., 28, 12711274, doi:10.1029/2000GL011418.

    • Search Google Scholar
    • Export Citation
  • Levitus, S., 1982: Climatological Atlas of the World Ocean. NOAA Prof. Paper 13, 173 pp. and 17 microfiche.

  • Liu, L. L., , and R. X. Huang, 2012: The global subduction/obduction rates: Their interannual and decadal variability. J. Climate, 25, 10961115, doi:10.1175/2011JCLI4228.1.

    • Search Google Scholar
    • Export Citation
  • Luo, Y., , L. M. Rothstein, , R.-H. Zhang, , and A. J. Busalacchi, 2005: On the connection between South Pacific subtropical spiciness anomalies and decadal equatorial variability in an ocean general circulation model. J. Geophys. Res., 110, C10002, doi:10.1029/2004JC002655.

    • Search Google Scholar
    • Export Citation
  • Mignot, J., , and C. Frankignoul, 2004: Interannual to interdecadal variability of sea surface salinity in the Atlantic and its link to the atmosphere in a coupled model. J. Geophys. Res., 109, C04005, doi:10.1029/2003JC002005.

    • Search Google Scholar
    • Export Citation
  • Niiler, P., 1975: Deepening of the wind-mixed layer. J. Mar. Res., 33, 405422.

  • O’Connor, B. M., , R. A. Fine, , and D. B. Olson, 2005: A global comparison of subtropical underwater formation rate. Deep-Sea Res., 52, 15691590, doi:10.1016/j.dsr.2005.01.011.

    • Search Google Scholar
    • Export Citation
  • Qiu, B., , and R. X. Huang, 1995: Ventilation of the North Atlantic and North Pacific: Subduction versus obduction. J. Phys. Oceanogr., 25, 23742390, doi:10.1175/1520-0485(1995)025<2374:VOTNAA>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Qu, T., , and J. Chen, 2009: A North Pacific decadal variability in subduction rate. Geophys. Res. Lett., 36, L22602, doi:10.1029/2009GL040914.

    • Search Google Scholar
    • Export Citation
  • Qu, T., , H. Mitsudera, , and T. Yamagata, 1999: A climatology of the circulation and water mass distribution near the Philippine coast. J. Phys. Oceanogr., 29, 14881505, doi:10.1175/1520-0485(1999)029<1488:ACOTCA>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Qu, T., , S. Gao, , and I. Fukumori, 2011: What governs the North Atlantic salinity maximum in a global GCM? Geophys. Res. Lett., 38, L07602, doi:10.1029/2011GL046757.

    • Search Google Scholar
    • Export Citation
  • Qu, T., , S. Gao, , and I. Fukumori, 2013: Formation of salinity maximum water and its contribution to the overturning circulation in the North Atlantic as revealed by a global general circulation model. J. Geophys. Res. Oceans, 118, 19821994, doi:10.1002/jgrc.20152.

    • Search Google Scholar
    • Export Citation
  • Reverdin, G., , E. Kestenare, , C. Frankignoul, , and T. Delcroix, 2007: Surface salinity in the Atlantic Ocean (30°S–50°N). Prog. Oceanogr., 73, 311340, doi:10.1016/j.pocean.2006.11.004.

    • Search Google Scholar
    • Export Citation
  • Ruddick, B., 1983: A practical indicator of stability of the water column to double-diffusive activity. Deep-Sea Res., 30A, 11051107, doi:10.1016/0198-0149(83)90063-8.

    • Search Google Scholar
    • Export Citation
  • Sasaki, Y. N., , N. Schneider, , N. Maximenko, , and K. Lebedev, 2010: Observational evidence for propagation of decadal spiciness anomalies in the North Pacific. Geophys. Res. Lett., 37, L07708, doi:10.1029/2010GL042716.

    • 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, doi:10.1175/1520-0485(1981)011<1015:FOTTSR>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Schmitt, R. W., 1999: Spice and the demon. Science, 283, 498499, doi:10.1126/science.283.5401.498.

  • Schneider, N., 2004: The response of tropical climate to the equatorial emergence of spiciness anomalies. J. Climate, 17, 10831095, doi:10.1175/1520-0442(2004)017<1083:TROTCT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Schneider, N., , A. J. Miller, , M. A. Alexander, , and C. Deser, 1999: Subduction of decadal North Pacific temperature anomalies: Observations and dynamics. J. Phys. Oceanogr., 29, 10561070, doi:10.1175/1520-0485(1999)029<1056:SODNPT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Schott, F., and et al. , 2003: The zonal currents and transports at 35°W in the tropical Atlantic. Geophys. Res. Lett., 30, 1349, doi:10.1029/2002GL016849.

    • Search Google Scholar
    • Export Citation
  • Stommel, H., 1979: Determination of watermass properties of water pumped down from the Ekman layer to the geostrophic flow below. Proc. Natl. Acad. Sci. USA, 76, 30513055, doi:10.1073/pnas.76.7.3051.

    • Search Google Scholar
    • Export Citation
  • Wallace, J. M., , C. Smith, , and Q. Jiang, 1990: Spatial patterns of atmosphere–ocean interaction in the northern winter. J. Climate, 3, 990998, doi:10.1175/1520-0442(1990)003<0990:SPOAOI>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Wang, C., , S. Dong, , and E. Munoz, 2010: Seawater density variations in the North Atlantic and the Atlantic meridional overturning circulation. Climate Dyn., 34, 953968, doi:10.1007/s00382-009-0560-5.

    • Search Google Scholar
    • Export Citation
  • Williams, R. G., 1991: The role of the mixed layer in setting the potential vorticity of the main thermocline. J. Phys. Oceanogr., 21, 18031814, doi:10.1175/1520-0485(1991)021<1803:TROTML>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Woods, J. D., 1985: The physics of pycnocline ventilation. Coupled Ocean-Atmosphere Models, J. C. J. Nihoul, Ed., Elsevier, 543–590.

  • Worthington, L. V., 1976: On the North Atlantic Circulation. Johns Hopkins University Press, 110 pp.

  • Yeager, S. G., , and W. G. Large, 2004: Late-winter generation of spiciness on subducted isopycnals. J. Phys. Oceanogr., 34, 15281547, doi:10.1175/1520-0485(2004)034<1528:LGOSOS>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Yeager, S. G., , and W. G. Large, 2007: Observational evidence of winter spice injection. J. Phys. Oceanogr., 37, 28952919, doi:10.1175/2007JPO3629.1.

    • Search Google Scholar
    • Export Citation
  • Zhang, D., , M. J. McPhaden, , and W. E. Johns, 2003: Observational evidence for flow between the subtropical and tropical Atlantic: The Atlantic subtropical cells. J. Phys. Oceanogr., 33, 17831797, doi:10.1175/2408.1.

    • Search Google Scholar
    • Export Citation
  • Zhang, L., , and T. Qu, 2014: Low-frequency variability of South Pacific Tropical Water from Argo. Geophys. Res. Lett., 41, 24412446, doi:10.1002/2014GL059490.

    • Search Google Scholar
    • Export Citation
Save