• Aoki, Y., , T. Suga, , and K. Hanawa, 2002: Subsurface subtropical fronts of the North Pacific as inherent boundaries in the ventilated thermocline. J. Phys. Oceanogr., 32 , 22992311.

    • Search Google Scholar
    • Export Citation
  • Bingham, F. M., 1992: Formation and spreading of Subtropical Mode Water in the North Pacific. J. Geophys. Res., 97 , 1117711189.

  • Conkright, M. E., and Coauthors, 1999: World Ocean Database 1998 documentation and quality control version 2.0. National Oceanographic Data Center Internal Rep. 14, 113 pp.

    • Search Google Scholar
    • Export Citation
  • Curry, R. G., 1996: HydroBase—A database of hydrographic stations and tools for climatological analysis. Woods Hole Oceanographic Institution Tech. Rep. WHOI-96-01, 44 pp.

    • Search Google Scholar
    • Export Citation
  • Deser, C. M., , A. Alexander, , and M. S. Timlin, 1996: Upper-ocean thermal variations in the North Pacific during 1970–1991. J. Climate, 9 , 18401855.

    • Search Google Scholar
    • Export Citation
  • Hanawa, K., , and I. Hoshino, 1988: Temperature structure and mixed layer over the Izu Ridge. J. Mar. Res., 46 , 683700.

  • Hanawa, K., , and L. D. Talley, 2001: Mode waters. Ocean Circulation and Climate, J. Church et al., Eds., Academic Press, 373–386.

  • Hautala, S. L., , and D. H. Roemmich, 1998: Subtropical mode water in the Northeast Pacific basin. J. Geophys. Res., 103 , 1305513066.

  • Hosoda, S., , S-P. Xie, , K. Takeuchi, , and M. Nonaka, 2001: Eastern North Pacific Subtropical Mode Water in a general circulation model: Formation mechanism and salinity effects. J. Geophys. Res., 106 , 1967119681.

    • Search Google Scholar
    • Export Citation
  • Huang, R. X., , and B. Qiu, 1994: Three-dimensional structure of the wind-driven circulation in the subtropical North Pacific. J. Phys. Oceanogr., 24 , 16081622.

    • Search Google Scholar
    • Export Citation
  • Huang, R. X., , and S. Russell, 1994: Ventilation of the subtropical North Pacific. J. Phys. Oceanogr., 24 , 25892605.

  • Johnson, G. C., , and M. J. McPhaden, 1999: Interior pycnocline flow from the subtropical to the equatorial Pacific Ocean. J. Phys. Oceanogr., 29 , 30733089.

    • Search Google Scholar
    • Export Citation
  • Kara, A. B., , P. A. Rochford, , and H. E. Hurlburt, 2000: An optimal definition for ocean mixed layer depth. J. Geophys. Res., 105 , 1680316821.

    • Search Google Scholar
    • Export Citation
  • Kawabe, M., , and K. Taira, 1998: Water masses and properties at 165°E in the western Pacific. J. Geophys. Res., 103 , 1294112958.

  • Kubokawa, A., 1999: Ventilated thermocline strongly affected by a deep mixed layer: A theory for subtropical countercurrent. J. Phys. Oceanogr., 29 , 13141333.

    • Search Google Scholar
    • Export Citation
  • Kubokawa, A., , and T. Inui, 1999: Subtropical countercurrent in an idealized ocean GCM. J. Phys. Oceanogr., 29 , 13031313.

  • Ladd, C. A., , and L. Thompson, 2000: Formation mechanisms for North Pacific central and eastern subtropical mode waters. J. Phys. Oceanogr., 30 , 868887.

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

  • Levitus, S., , and T. P. Boyer, 1994: Temperature. Vol. 4, World Ocean Atlas 1994, NOAA Atlas NESDIS 4, 111 pp.

  • Levitus, S., , R. Burgett, , and T. P. Boyer, 1994: Salinity. Vol. 3, World Ocean Atlas 1994, NOAA Atlas NESDIS 3, 111 pp.

  • Lozier, S. M., , W. B. Owens, , and R. G. Curry, 1995: The climatology of the North Atlantic. Progress in Oceanography, Vol. 36, Pergamon, 1–44.

    • Search Google Scholar
    • Export Citation
  • Luyten, J. R., , J. Pedlosky, , and H. Stommel, 1983: The ventilated thermocline. J. Phys. Oceanogr., 13 , 292309.

  • Macdonald, A. M., , T. Suga, , and R. G. Curry, 2001: An isopycnally averaged North Pacific climatology. J. Atmos. Oceanic Technol., 18 , 394420.

    • Search Google Scholar
    • Export Citation
  • Marshall, J. C., , A. J. G. Nurser, , and R. G. Williams, 1993: Inferring the subduction rate and period over the North Atlantic. J. Phys. Oceanogr., 23 , 13151329.

    • Search Google Scholar
    • Export Citation
  • McCartney, M. S., 1982: The subtropical recirculation of mode waters. J. Mar. Res., 40 , (Suppl.),. 427464.

  • Mecking, S., , and M. J. Warner, 2001: On the subsurface CFC maxima in the subtropical North Pacific thermocline and their relation to mode waters and oxygen maxima. J. Geophys. Res., 106 , 2217922198.

    • Search Google Scholar
    • Export Citation
  • Nakamura, H., 1996: A pycnostad on the bottom of the ventilated portion in the central subtropical North Pacific: Its distribution and formation. J. Oceanogr., 52 , 171188.

    • Search Google Scholar
    • Export Citation
  • NODC, 1994: World Ocean Atlas 1994 CD-ROM data set documentation. National Oceanographic Data Center Informal Rep. 13, 30 pp.

  • Qu, T., , H. Mitsudera, , and B. Qiu, 2001: A climatological view of the Kuroshio/Oyashio system east of Japan. J. Phys. Oceanogr., 31 , 25752589.

    • Search Google Scholar
    • Export Citation
  • Rhines, P. B., , and W. R. Young, 1982: A theory of the wind-driven circulation. I. Mid-ocean gyre. J. Mar. Res, 40 , (Suppl.),. 559596.

    • Search Google Scholar
    • Export Citation
  • Roden, G. I., 1970: Aspects of the mid-Pacific transition zone. J. Geophys. Res., 75 , 10971109.

  • 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.

    • Search Google Scholar
    • Export Citation
  • Smith, W. H. F., , and P. Wessel, 1990: Gridding with continuous curvature splines in tension. Geophysics, 55 , 293305.

  • Stommel, H. M., 1979: Determination of water mass properties of water pumped down from Ekman layer to the geostrophic flow below. Proc. Natl. Acad. Sci. USA, 76 , 30513055.

    • Search Google Scholar
    • Export Citation
  • Suga, T., , and K. Hanawa, 1990: The mixed layer climatology in the northwestern part of the North Pacific subtropical gyre and the formation area of Subtropical Mode Water. J. Mar. Res., 48 , 543566.

    • Search Google Scholar
    • Export Citation
  • Suga, T., , and K. Hanawa, 1995: The subtropical mode water circulation in the North Pacific. J. Phys. Oceanogr., 25 , 958970.

  • Suga, T., , K. Hanawa, , and Y. Toba, 1989: Subtropical mode water in the 137°E section. J. Phys. Oceanogr., 19 , 16051618.

  • Suga, T., , Y. Takei, , and K. Hanawa, 1997: Thermostad distribution in the North Pacific subtropical gyre: The central mode water and the subtropical mode water. J. Phys. Oceanogr., 27 , 140152.

    • Search Google Scholar
    • Export Citation
  • Suga, T., , A. Kato, , and K. Hanawa, 2000: North Pacific Tropical Water: Its climatology and temporal changes associated with the climate regime shift in the 1970s. Progress in Oceanography, Vol. 47, Pergamon, 223–256.

    • Search Google Scholar
    • Export Citation
  • Talley, L. D., 1985: Ventilation of the subtropical North Pacific: The shallow salinity minimum. J. Phys. Oceanogr., 15 , 633649.

  • Talley, L. D., 1988: Potential vorticity distribution in the North Pacific. J. Phys. Oceanogr., 18 , 89106.

  • Talley, L. D., 1991: An Okhotsk Sea water anomaly: Implications for ventilation in the North Pacific. Deep-Sea Res., 38 , (Suppl.),. S171S190.

    • Search Google Scholar
    • Export Citation
  • Talley, L. D., 1993: Distribution and formation of the North Pacific Intermediate Water. J. Phys. Oceanogr., 23 , 517537.

  • Williams, R. G., 1989: The influence of air–sea interaction on the ventilated thermocline. J. Phys. Oceanogr., 19 , 12551267.

  • Williams, R. G., 1991: The role of the mixed layer in setting the potential vorticity of the main thermocline. J. Phys. Oceanogr., 21 , 18031814.

    • Search Google Scholar
    • Export Citation
  • Xie, S-P., , T. Kunitani, , A. Kubokawa, , M. Nonaka, , and S. Hosoda, 2000: Interdecadal thermocline variability in the North Pacific for 1958–97: A GCM simulation. J. Phys. Oceanogr., 30 , 27982813.

    • Search Google Scholar
    • Export Citation
  • Yasuda, I., 1997: The origin of the North Pacific Intermediate Water. J. Geophys. Res., 102 , 893909.

  • Young, W. R., , and P. B. Rhines, 1982: A theory of wind-driven circulation. II: Gyre with western boundary layers. J. Mar. Res., 40 , 849872.

    • Search Google Scholar
    • Export Citation
  • Yuan, X., , and L. D. Talley, 1992: Shallow salinity minima in the North Pacific. J. Phys. Oceanogr., 22 , 13021316.

  • Yuan, X., , and L. D. Talley, 1996: The subarctic frontal zone in the North Pacific: Characteristics of frontal structure from climatological data and synoptic surveys. J. Geophys. Res., 101 , 1649116508.

    • Search Google Scholar
    • Export Citation
  • Zhang, H-M., , and N. G. Hogg, 1992: Circulation and water mass balance in the Brazil Basin. J. Mar. Res., 50 , 385420.

  • View in gallery

    The station positions of (a) the dataset used for the annual mean isopycnal climatology and (b) the dataset used for the late winter (Feb–Mar) NWMLC

  • View in gallery

    The distribution of (a) the number of stations and (b) the search radius at each grid point for the late winter (Feb–Mar) NWMLC

  • View in gallery

    The late winter (Feb–Mar) climatology of the (left) average and (right) standard deviation of (a), (b) mixed layer depth (contour interval of 25 m/10 m with each 50 m/20 m thickened, respectively), (c), (d) sea surface density (contour interval of 0.2 σθ/0.1 σθ with each 1.0 σθ/0.2 σθ thickened, respectively), (e), (f) sea surface temperature (contour interval of 1°C/0.5°C with each 5°C/1°C thickened, respectively), and (g), (h) sea surface salinity (contour interval of 0.1 psu/0.1 psu with each 0.5 psu/0.2 psu thickened, respectively)

  • View in gallery

    (Continued)

  • View in gallery

    (a) WOA94 mixed layer depth (contour interval of 25 m with each 50 m thickened) in Mar and (b) its difference (contour interval of 20 m with each 40 m thickened) from the NWMLC [present study (P.S.)]. (c) The WOA94 sea surface density (contour interval of 0.2 σθ with each 1.0 σθ thickened) in Mar and (d) its difference (contour interval of 0.1 σθ with each 0.2 σθ thickened) from the NWMLC (P.S.). The dashed contours denote negative differences

  • View in gallery

    The annual mean climatology of Q (10−12 m−1 s−1) on the isopycnal surfaces (a) 24.2 σθ, (b) 24.6 σθ, (c) 25.0 σθ, (d) 25.2 σθ, (e) 25.4 σθ, (f) 25.6 σθ, (g) 25.8 σθ, (h) 26.0 σθ, (i) 26.2 σθ, (j) 26.4 σθ, (k) 26.5 σθ, (l) 26.6 σθ, (m) 26.7 σθ, (n) 26.8 σθ, (o) 26.9 σθ, and (p) 27.0 σθ. The contour interval is 100 × 10−12 m−1 s−1 with each 200 × 10−12 m−1 s−1 thickened. Additional dashed contours at intervals of 25 × 10−12 m−1 s−1 are drawn for Q lower than 300 × 10−12 m−1 s−1. Thick white contours denote the winter outcrop of each isopycnal. Thin white contours represent pressure anomaly streamfunction (contour interval of 1 m2 s−2) relative to 2000 dbar. The region where the isopycnal is defined is shaded. The darkest, the second darkest, and the third darkest shades denote the regions with Q lower than 200 × 10−12 m−1 s−1, from 200 × 10−12 to 300 × 10−12 m−1 s−1, and from 300 × 10−12 to 500 × 10−12 m−1 s−1, respectively

  • View in gallery

    (Continued)

  • View in gallery

    (Continued)

  • View in gallery

    (Continued)

  • View in gallery

    Potential temperature–salinity diagram for the low-Q water representing mode water cores. Circles, squares, and triangles represent the STMW, CMW, and ESTMW cores, respectively

  • View in gallery

    Grid points in the NWMLC, which have properties corresponding to those of the mode water cores as defined in Fig. 6, superimposed on the maps of (a) the sea surface density (contour interval of 0.2 σθ with each 1.0 σθ thickened) and (b) the mixed layer depth (contour interval of 25 m with each 100 m thickened) in the late winter. Circles, squares, and triangles represent the STMW, CMW, and ESTMW cores, respectively. The annual mean climatology of the geopotential anomaly (contour interval of 1 m2 s−2) at 200 dbar relative to 2000 dbar is also superimposed on both maps

  • View in gallery

    Grid points in the WOA94 Mar mixed layer climatology, which have properties corresponding with those of the mode water cores as defined in Fig. 6, superimposed on the maps of (a) the sea surface density and (b) the mixed layer depth in Mar. Circles, squares, and triangles represent the STMW, CMW, and ESTMW cores, respectively

  • View in gallery

    Schematics showing the terms determining Q of the fluid subducted from the bottom of steady mixed layer: cross-isopycnal flow, −uρm; lateral induction, −uh; and vertical pumping, −w

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 290 290 20
PDF Downloads 161 161 17

The North Pacific Climatology of Winter Mixed Layer and Mode Waters

View More View Less
  • 1 Department of Geophysics, Graduate School of Science, Tohoku University, Sendai, Japan
  • | 2 Woods Hole Oceanographic Institution, Woods Hole, Massachusetts
© Get Permissions
Full access

Abstract

A climatology of the winter mixed layer in the North Pacific Ocean was constructed using hydrographic data from historical archives and recent observational programs, including the World Ocean Circulation Experiment. The main aim was to provide better knowledge about source areas of upper water masses. The authors have endeavored to preserve water properties near the frontal regions by keeping the smoothing scale as small as possible. The resulting climatology shows considerable differences in the mixed layer depth and its water properties from those derived from the World Ocean Atlas (WOA). Maps of the potential vorticity field of the North Pacific pycnocline are presented using the isopycnally averaged climatology, HydroBase. Three distinct lateral minima of potential vorticity are identified as Subtropical Mode Water (STMW), Central Mode Water (CMW), and Eastern Subtropical Mode Water (ESTMW), in the western, central, and eastern parts of the subtropical gyre, respectively. The HydroBase isopycnal climatology is more consistent with the present mixed layer climatology than with the mixed layer from WOA in the sense that the former represents the formation of all mode waters more adequately. The STMW and ESTMW formation areas are associated with the mixed layer front and the small horizontal gradient of the mixed layer density, respectively, which confirms previously proposed formation mechanisms. That is, the low potential vorticity of STMW and ESTMW results from the large lateral induction and the small cross-isopycnal flow, respectively. The CMW formation area is not primarily associated with the mixed layer front, which contrasts with previous ideas. It is suggested that low potential vorticity of CMW is mainly caused by small cross-isopycnal flow rather than through large lateral induction rate. Additional new features of subtropical pycnocline ventilation revealed by the HydroBase isopycnal climatology are also discussed.

Corresponding author address: Dr. Toshio Suga, Department of Geophysics, Graduate School of Science, Tohoku University, Aoba-ku, Sendai 980-8578, Japan. Email: suga@pol.geophys.tohoku.ac.jp

Abstract

A climatology of the winter mixed layer in the North Pacific Ocean was constructed using hydrographic data from historical archives and recent observational programs, including the World Ocean Circulation Experiment. The main aim was to provide better knowledge about source areas of upper water masses. The authors have endeavored to preserve water properties near the frontal regions by keeping the smoothing scale as small as possible. The resulting climatology shows considerable differences in the mixed layer depth and its water properties from those derived from the World Ocean Atlas (WOA). Maps of the potential vorticity field of the North Pacific pycnocline are presented using the isopycnally averaged climatology, HydroBase. Three distinct lateral minima of potential vorticity are identified as Subtropical Mode Water (STMW), Central Mode Water (CMW), and Eastern Subtropical Mode Water (ESTMW), in the western, central, and eastern parts of the subtropical gyre, respectively. The HydroBase isopycnal climatology is more consistent with the present mixed layer climatology than with the mixed layer from WOA in the sense that the former represents the formation of all mode waters more adequately. The STMW and ESTMW formation areas are associated with the mixed layer front and the small horizontal gradient of the mixed layer density, respectively, which confirms previously proposed formation mechanisms. That is, the low potential vorticity of STMW and ESTMW results from the large lateral induction and the small cross-isopycnal flow, respectively. The CMW formation area is not primarily associated with the mixed layer front, which contrasts with previous ideas. It is suggested that low potential vorticity of CMW is mainly caused by small cross-isopycnal flow rather than through large lateral induction rate. Additional new features of subtropical pycnocline ventilation revealed by the HydroBase isopycnal climatology are also discussed.

Corresponding author address: Dr. Toshio Suga, Department of Geophysics, Graduate School of Science, Tohoku University, Aoba-ku, Sendai 980-8578, Japan. Email: suga@pol.geophys.tohoku.ac.jp

1. Introduction

The winter mixed layer produces the densest water at a given locale and provides source water types of various water masses. A good example of this situation is the widely accepted concept that only the fluid leaving the late winter mixed layer can be irreversibly subducted into the subtropical permanent pycnocline (Stommel 1979). That is, the water masses in the ventilated pycnocline originate from the winter mixed layer. Formation mechanisms and variability of those water masses cannot be understood without proper knowledge about the winter mixed layer such as the spatial distribution of its thickness and water properties.

In the North Pacific, many studies addressing mean properties and/or variability of the water masses in the ventilated pycnocline (e.g., Yuan and Talley 1992; Huang and Qiu 1994; Deser et al. 1996; Hautala and Roemmich 1998; Ladd and Thompson 2000; Suga et al. 2000) have drawn information about the winter mixed layer, or the winter outcrops of isopycnals, from a series of climatological atlases known as World Ocean Atlases (WOAs; Levitus 1982; Levitus and Boyer 1994; Levitus et al. 1994). Although the globally consistent information about the winter mixed layer provided by WOAs is useful in many cases, it is not always the most appropriate information. For example, North Pacific Subtropical Mode Water (STMW) and North Pacific Central Mode Water (CMW) are formed near the Kuroshio Front and the Kuroshio Bifurcation Front, respectively (Hanawa and Hoshino 1988; Suga and Hanawa 1990; Bingham 1992; Nakamura 1996; Suga et al. 1997). The WOAs may not capture their formation areas near the fronts adequately because of their largely smoothed nature.

The detailed mean spatial structure and properties of the STMW formation area have been discussed based on elaborate analyses of the winter mixed layer in which mean fields were calculated with respect to stream coordinate systems (e.g., Hanawa and Hoshino 1988; Bingham 1992). The stream coordinate systems use the positions of the fronts, or the axes of the currents, as their origins and thus have the advantage of preserving the frontal structures in the mean fields better. They are useful for describing formation areas of any water masses found near the fronts. They require, however, information about station positions relative to the relevant fronts, which is not always available and thus limits the application of the technique. Another limitation of the stream coordinate systems is that resultant mean fields are described relative to particular fronts and cannot easily be placed geographically with a basinwide framework.

Recently, Macdonald et al. (2001) presented a new hydrographic climatology for the North Pacific, intending to provide the mean water mass properties and circulation with finer resolution. The climatology was constructed by isopycnal averaging with the smallest possible smoothing scales. It is an extension of the similar climatology for the North Atlantic known as HydroBase (Lozier et al. 1995; Curry 1996) and thus called the North Pacific HydroBase (henceforth NPHB). It was shown that the NPHB preserved a number of features in water property distribution better than the previous climatologies especially near the coastal boundaries and the frontal regions.

While the wintertime outcrops of isopycnals are available through an algorithm provided by HydroBase (Curry 1996), all of the winter mixed layer properties, including its thickness, are not directly provided by HydroBase itself. To utilize NPHB fully for describing formation and circulation of various water masses, we need a winter mixed layer climatology fitted to the NPHB. The purpose of the present paper is to present such a climatology, one that has better resolution in both space and temperature–salinity plane and thus a better connection to NPHB than existing climatologies. The new winter mixed layer climatology (henceforth NWMLC) is then used along with the NPHB isopycnal maps to examine formation and spreading of mode waters that are good tracers of subtropical ventilation (Hanawa and Talley 2001) and so are suitable for testing the consistency between the NWMLC and the NPHB climatology.

The rest of this paper is organized as follows. Section 2 describes the hydrographic dataset used to construct the climatology, the definition of the mixed layer, and the procedure of averaging and gridding. Property fields of the NWMLC are described in section 3. Comparison between the NWMLC and the winter mixed layer derived from the WOA is made and the major differences are described. In section 4, the features of mode waters are highlighted in the NPHB isopycnal climatology and compared with the properties of the NWMLC, which leads to identification of plausible formation areas of the mode waters. Implications from the NWMLC about the formation mechanisms of the mode waters and those from the NPHB isopycnal climatology about the ventilation of the overall subtropical pycnocline are discussed in section 5.

2. Data and processing

The major hydrographic dataset for this study comes from the NPHB produced by Macdonald et al. (2001), which is then supplemented by several additional sources. The NPHB dataset consists of two parts. The first part is the observed data, some 115 000 profiles, of World Ocean Atlas 1994 (henceforth WOA94; NODC 1994), which have been passed through the quality control procedure mainly consisting of the statistical check of temperature (T) and salinity (S) on the TS diagram. The second part is the CTD data from several World Ocean Circulation Experiment (WOCE) sections and pre-WOCE sections. The additions to this original NPHB dataset are the other WOCE sections available as of January 2001 and the CTD data collected by National Oceanic and Atmospheric Administration/Pacific Marine Environmental Laboratory (NOAA/PMEL; Johnson and McPhaden 1999). The horizontal coverage of the whole dataset is illustrated in Fig. 1a. The isopycnal climatology discussed in this study was calculated from this updated NPHB.

Since the mixed layer properties at the end of winter are most relevant to the sources of many water masses in the permanent pycnocline, the mixed layer climatology was constructed over the two late winter months, February and March. As noted by Macdonald et al. (2001), the data distribution in the winter months is extremely limited. We found several blank areas, lacking in data for February and March, which could have greatly affected the gridded climatology of the winter mixed layer. To reduce this problem, additional data from World Ocean Database 1998, version 2 (henceforth WOD98; Conkright et al. 1999), were included in the data-sparse regions. These newer data have passed through all the quality control procedures run by National Oceanographic Data Center (NODC) and our final check provided by a subjective visual scanning of areal property maps as done by Macdonald et al. (2001). The station distribution of the February–March dataset is displayed in Fig. 1b.

The mixed layer depth was defined as the depth at which potential density is different from the sea surface density by 0.125 σθ. This simple definition is the same as adopted by numerous previous studies (e.g., Levitus 1982; Huang and Qiu 1994; Ladd and Thompson 2000). Although the criterion of 0.125 σθ is arbitrary, we have confirmed that the resulting mixed layer depth was not particularly sensitive to a criterion ranging from 0.075 σθ to 0.15 σθ with differences less than 5%–15%. Since it is known that the quality of the sea surface data are generally poor due to bucket sampling, we eliminated all the sea surface data and regarded the data immediately below the sea surface, nominally 10 m in depth, as the sea surface data. The mixed layer depth was derived by linearly interpolating the observed data for individual profiles. The mixed layer averaged properties, including σθ, potential temperature (θ) and salinity, were also calculated for individual profiles.

Using this unevenly distributed set of the mixed layer properties at individual stations, the next step was to create a 1° gridded dataset of the climatology. Our main concern in producing the climatology was to preserve actual water properties in the mixed layer as much as possible. A similar goal had been achieved for the HydroBase climatology by adopting the averaging and smoothing procedure characterized by isopycnal averaging and small smoothing scale (Lozier et al. 1995; Macdonald et al. 2001). The isopycnal averaging is effective to avoid smoothing over different isopycnals, which could create artificial water properties. However, this technique does not work in the present case. Since the isopycnals are virtually vertical in the mixed layer, any horizontal averaging procedures will involve artificial “diapycnal mixing” (Curry 1996). We therefore followed only the latter part of the previous averaging and smoothing scheme. That is, we tried to keep the smoothing scale as small as possible.

A gridded matrix in which each grid point was centered on a 1° square was established for the geographical domain. Mean properties at each grid point were evaluated by a low-pass filter based on the density of observations per 1° square. In all gridding it was attempted to maintain at least 10 observations per bin by allowing a gradually increased search radius (d) with increments of 1°. Points outside the 1° grid block were distance weighted by a factor of d−2. This technique results in lesser smoothing in the well sampled regions and greater smoothing in the more poorly sampled regions. We adopted this fairly simple method as the first attempt to produce a mixed layer climatology with finer resolution. Although more sophisticated schemes may be applied in future study, even the present simple scheme provided a NWMLC that was fairly consistent with the NPHB isopycnal climatology in terms of water mass properties as shown in the subsequent sections.

The number of stations and the search radius used at each grid point are shown in Fig. 2. The well-sampled regions near the coasts of Japan and North America, and around several interior sites of oceanographic stations, show large amounts of data and small search radii. More than 10 observations are maintained at most of the grid points over the North Pacific by allowing a search radius within 3° except for the limited areas centered at 30°N, 185°E; at 15°N, 180°; and at 10°N, 225°E.

To produce contours, a 2° block mean was applied to reduce the grid-scale variability, followed by a final smoothing with the mapping algorithm of Smith and Wessel (1990), which fits a surface with continuous second derivatives and minimal curvature to the gridded data. This last procedure is the same as that used by Lozier et al. (1995) and Macdonald et al. (2001).

3. Property fields of the winter mixed layer climatology

The NWMLC maps of the mixed layer depth (MLD) and the sea surface density (SSD), temperature (SST), and salinity (SSS) are presented in Fig. 3. The MLD map shows many spatial features north of about 25°N while it is fairly uniform ranging from only 25 to 75 m in the southern latitudes. Within the open ocean, there are two MLD maxima deeper than 200 m along 32° and 42°N near the western boundary. The two maxima are embedded in a broad band of fairly deep mixed layer represented by the 150-m contour extending eastward from the western boundary to almost 200°E.

The basinwide distribution of SSD is characterized by a much larger horizontal gradient in the subtropical zone between 10° and 40°N than in the northern and southern latitudes. This is especially true west of 200°E. The SSD gradient is predominantly northward in this subtropical zone while the SSD gradient north of 40°N is northwestward or westward and almost an order of magnitude weaker. A band of subtropical high SSD gradient branches off near 35°N, 200°E extending northeastward into the Gulf of Alaska and northwestward along the Canadian and Alaskan coast. As a result, the SSD gradient north of 40°N is considerably higher east of 200°E than that west of it. The branching of the high SSD gradient leaves a region of fairly weak SSD gradient at 25°–35°N, 210°–240°E.

Throughout the basin the SST distribution is more zonal than the SSD distribution. The SST gradient is mostly southward and larger at 20°–45°N than that both south and north of this zone. The SST gradient north of 45°N is southeastward, with the lowest SST near Kamchatka; that at 25°–35°N in the eastern basin east of 210°E is southwestward. It is noteworthy that considerable SST gradient is found at 40°–45°N west of 200°E where SSD is fairly uniform. The SST gradient is also quite large in the eastern area of the weak density gradient at 25°–35°N east of 210°E mentioned above.

The distribution of SSS is characterized by a few large-scale extrema in contrast with those of SSD and SST, which vary more or less monotonically from the south to the north. The subarctic minimum extends from the Canadian and Alaskan coast both westward and southward. The subtropical maximum extends zonally almost along the tropic of Cancer. Immediately south of it is the tropical minimum associated with the intertropical convergence zone. Another maximum appears on the equator 160°–230°E, which is actually the northern margin of the high SSS region extending from the subtropics of the South Pacific. These extrema are separated by fronts. The sharpest front extends eastward along about 40°N from the Japanese coast to 200°E, and then turns southeastward directly bounding the low SSS extending from the north and the high SSS of the subtropical maximum. This front is the Subarctic Front, or the subarctic frontal zone described by Yuan and Talley (1996). The large SSS gradient of this front largely compensates the SST gradient to yield the areas of weak SSD gradient described above.

In the next section, we examine the properties of the NWMLC with respect to their correspondence to water properties of the permanent pycnocline in the NPHB isopycnal climatology, especially focusing on the mode waters. Before we proceed, a brief comparison is made between the NWMLC and the winter mixed layer derived from WOA94. The comparison is useful because the latter has been widely used by a number of previous studies (e.g., Kara et al. 2000; Ladd and Thompson 2000; Suga et al. 2000).

Figure 4 shows the maps of MLD derived from the vertical σθ profile at each 1° grid point and SSD in March from WOA94 and the differences of those fields from the NWMLC. The overall distribution of the MLD and SSD derived from WOA94 is similar to that derived from the NWMLC. However, each of the difference maps shows recognizable systematic patterns. The largest differences in MLD are found near the two MLD maxima off the Japanese coast. Although the two maxima also appear in the WOA94 map, the southern maximum is diminished and the northern one enhanced over those found in NWMLC. The WOA94 SSD is considerably higher, by 0.1–0.2 σθ than the NWMLC SSD in the band along about 40°N, which overlaps with the WOA94 enhanced northern MLD maximum. On the other hand, the WOA94 SSD is lower than the NWMLC SSD in the area south of 30°–35°N where the MLD from the WOA94 is considerably shallower than the MLD of NWMLC.

The causes of these differences are not immediately apparent because the two mixed layer climatologies differ in several aspects. While the NWMLC is derived from the data for February and March, the WOA94 climatology is based solely on the March data. For comparison, we calculated the WOA94 climatology averaged over February and March and their differences from NWMLC (not shown). The difference fields show essentially the same features at the two MLD maxima as those described above (Figs. 4b and 4d), while the MLD difference at the northern maximum is diminished by 20–40 m and that at the southern maximum is enhanced by about 20 m. Therefore, it is unlikely that the discrepancy in the definition of the late winter months is the primary cause of the differences between NWMLC and the WOA94 climatology.

Another obvious distinction is that the types of data used to calculate the climatology are not identical. The major source of the data for the NWMLC comes from the bottle and CTD data that were used to create the WOA94 climatology, which is then supplemented by other sources including the WOCE sections, NOAA/PMEL CTD data, and part of WOD98. On the other hand, the WOA94 climatology is based on not only the bottle and CTD data, but also mechanical bathythermograph (MBT) and expendable bathythermograph (XBT) data. Accordingly, the number of the temperature data utilized in WOA94 is several times as large as the number of salinity data. Only the stations that have both temperature and salinity data are used in the NWMLC. It is not clear, however, if this distinction is a major cause of the differences found in Figs. 4b and 4d.

There is one more distinction between the two climatologies, which is associated with the averaging procedures used to produce gridded climatology. In the NWMLC, the mixed layer properties and the sea surface properties are evaluated at individual stations and then averaged over typically the 1° to a few degrees square depending on the density of the data distribution. In the WOA94, temperature and salinity at each standard depth are independently gridded with the objective mapping procedure, which effectively smoothes the fields over the length scale of about 1000 km. A σθ profile at each grid point is then calculated from the corresponding gridded temperature and salinity profiles. The MLD is finally evaluated from that σθ profile. These differences in the data sources and gridding schemes presumably affect the resulting fields.

In light of the above distinction, one cause of the major differences in the two fields may be described as follows. In the NWMLC along the zone near 40°N, especially at 150°E–180°, SSD is significantly lower and MLD shallower than in WOA94. This zone, known as the subarctic frontal zone, corresponds to a large meridional gradient of temperature and salinity (e.g., Yuan and Talley 1996). The frontal zone is characterized by colder (warmer) and fresher (saltier) water to the north (south) with temperature and salinity gradients that are density compensating, as illustrated in Fig. 3. Smoothing the temperature and salinity fields across this front would result in artificially denser water in the frontal zone due to the nonlinear nature of the equation of state of the sea water. Since the front is confined to the top couple of hundred meters, the subsurface density would be much less affected by the smoothing, which would result in a larger MLD under the present definition. As a quick test for this idea, we calculated the alternative SSD field from gridded NWMLC SST and SSS (Figs. 3e,g). The resulting SSD is denser by 0.02–0.06 σθ than the original NWMLC SSD along the zone near 40°N at 150°E–180° (not shown). The results support the idea that a difference in averaging procedures can, at least partly, explain the large differences in the northern MLD maximum between the NWMLC and the mixed layer derived from WOA94.

The same idea is not applicable to the large differences in the southern MLD maximum along 30°N at 140°–160°E because here, there is no density-compensating temperature and salinity front. Instead, the NWMLC maps (Figs. 3a,c) show that the southern MLD maximum corresponds with the area of low SSD gradient between the Kuroshio Front along about 35°N and the Subtropical Front along about 25°N. These two fronts are considerably smoothed out and the low gradient area diminishes substantially in the SSD map of WOA94 (Fig. 4c). It is thus inferred that the southern MLD maximum is weakened in the WOA94 climatology because the smoothing scale is too large to retain this spatially confined feature.

There are likely a variety of other causes contributing to the differences between the NWMLC and the mixed layer derived from the WOA94. However, we believe that they are secondary compared to those discussed above. It is beyond the scope of the present study to investigate further the causes of differences. We will concentrate on deriving new information about source regions of upper water masses from NWMLC.

4. Connection between mode waters and the winter mixed layer

a. Mode waters in the potential vorticity maps

Mode waters are characterized by the vertical homogeneity of their water properties and thus identified as waters of low potential vorticity (Q) (e.g., McCartney 1982; Talley 1988; Suga et al. 1989). Maps of Q on the isopycnal surfaces from 24.0 σθ to 27.0 σθ by 0.1 σθ were produced using the annual mean climatology of NPHB. Here Q was derived neglecting the contribution of relative vorticity, which is an appropriate simplification except in regions of strong flows (Talley 1988), and Q was computed for each 1° gridded profile by estimating a vertical density gradient over the span of 100 dbar. A subset of the maps is shown in Fig. 5. As described below, three types of mode waters are depicted as distinct cores of lateral Q minima on the isopycnal maps: STMW, CMW, and North Pacific Eastern Subtropical Mode Water (ESTMW; Hautala and Roemmich 1998). Superimposed on the Q maps are the late winter isopycnal outcrops from the NWMLC and the annual mean pressure anomaly streamfunction (Zhang and Hogg 1992) relative to 2000 dbar. Talley (1988) presented similar isopycnal Q maps based on the earliest version of WOA (Levitus 1982) in a comprehensive work on the North Pacific Q distribution and its interpretation in the context of ventilation and ocean dynamics. Talley's maps captured the basin-scale features of the low Q regions corresponding to the three mode waters, but CMW and ESTMW were not fully recognized at that time and so were not discussed as distinct water masses. With isopycnal averaging, smaller-scale smoothing and an updated dataset, the present maps now reveal some more detailed characteristics of the Q fields associated with the mode waters.

STMW is identified as the lateral Q minimum near the northwestern corner of the subtropical gyre at 24.8–25.7 σθ surfaces. It has been shown that the spatial distribution of STMW is tightly associated with the anticyclonic recirculation of the Kuroshio (Suga and Hanawa 1995). The present maps demonstrate how the low Q signature of STMW, including its core and surrounding area, tends to fill the recirculation region delineated with one of closed streamlines, forming a low Q pool with the minimum at the center (see, e.g., Fig. 5f). Described below is how this feature contrasts with the distribution pattern of ESTMW and CMW.

The lateral Q minimum of ESTMW is obvious near 25°N and 220°E at the 24.2–25.3 σθ surfaces. The density range is slightly wider than that mentioned by Talley (1988). The spatial pattern of the low Q region resembles that appearing on the Q maps presented by Hautala and Roemmich (1998) using the Tropical Ocean and Global Atmosphere (TOGA) XBT data, which are totally independent of the present data source. The low Q, ESTMW core, is located in the south-southwestward flow near the eastern end of the subtropical gyre. The core is embedded in a corridor of low Q lying between the highs of the seasonal pycnocline and the Tropics as pointed out by Talley (1988). Although Talley inferred that this zonally elongated region of low Q was ventilated in the east near the ESTMW core, the streamlines overlaid upon the present maps suggest that the corridor is more likely to be ventilated directly to the north. The ESTMW core appears to be dissipated southward rapidly along the streamlines.

CMW appears as the isolated lateral Q minimum near the center of the North Pacific around 30°–40°N, 180° at 25.9–26.5 σθ surfaces. While the low Q core of STMW appeared to migrate successively to the east with increasing density up to the density range of CMW in the earlier maps (Talley 1988), the Q minimum of CMW is distinctly separated from that of STMW as suggested by Nakamura (1996) who showed a few isopycnal Q maps based on the Levitus (1982) climatology. The separation of the two mode waters is more clearly demonstrated here in the Q map at 25.8 σθ (Fig. 5g), where weak, but discernible, separate double minima exist. The western minimum centered at 25°N and 150°E marks the lower bounds of STMW and the eastern one centered at 28°N and 180° marks the upper bounds of CMW.

While the low Q signatures of STMW and ESTMW are much the same as those in the earlier maps, the low Q signature of CMW is somewhat different from that in the earlier maps. The lateral CMW Q minimum seen here penetrating down to the 26.5 σθ surface, reached only to the 26.2 σθ surface in the earlier maps (Talley 1988). The core of CMW is located in the eastward or southeastward flow of the subtropical gyre. The extension of the low Q tongue, or the low Q corridor, reaching southwestward from the core to the tropical western boundary at the 25.9–26.4 σθ surfaces was much less obvious in the earlier maps. The low Q tongue is less obvious at the 26.5 σθ surface because it is merged with the tropical low Q. Since the low Q tongue coincides with the streamlines, it is reasonable to suppose that the tongue primarily reflects southwestward advective spreading of CMW.

b. Probable formation sites of mode waters

Since the mode waters originate from the corresponding isopycnal outcrops, their low Q signature should emanate from somewhere along the wintertime outcrops. However, the correspondence between the low Q cores and surface outcrops is somewhat marred in the present maps as it was in earlier efforts (Talley 1988). For example, the low Q cores of STMW at 25.0–25.6 σθ, CMW at 26.2–26.5 σθ, and ESTMW at 25.2 σθ are not directly connected to the isopycnal outcrops but located several hundreds of kilometers south/southwest, east/southeast, and southeast of the nearest segments of the isopycnal outcrops, respectively. Although we employed fairly small scales of lateral smoothing compared to those in the WOAs, the smoothing near isopycnal outcrops has, nevertheless, produced these unsatisfactory results because we averaged the data over different seasons to compute annual means.

Consider a synoptic Q field in summer. The Q distribution at a given isopycnal must abruptly change across its wintertime outcrop. That is, the equatorward side of the isopycnal is in the deep permanent pycnocline and has fairly low Q, while its poleward side is in the shallow seasonal pycnocline and has considerably higher Q. Consequently the isopycnal averaging across the location of the wintertime outcrop will cause a broadened front of Q, which will tend to mar the connection between the isopycnal outcrop and the low Q signature. Actually, the dataset from which we computed the annual means has lots of summer stations; there are twice as many summer stations available as winter ones (Macdonald et al. 2001). Therefore, our annual mean maps are likely unable to produce clear signatures of low Q near the wintertime outcrops.

Seasonal mean maps would give better pictures near the wintertime outcrops as demonstrated by Suga and Hanawa (1995). They illustrated the low Q signatures of STMW connected to the isopycnal outcrops in the winter-mean climatological maps of the northwestern subtropical North Pacific. Although we tried to prepare seasonal mean maps over the North Pacific, the temporally and spatially uneven data distribution yielded considerably distorted fields. Therefore, for the purposes of the present study we concentrate our description and discussion on the annual mean fields.

Any properties that change abruptly across the wintertime outcrop suffer from the same problem. Streamfunctions are also affected because the depth of the isopycnal changes abruptly across the wintertime outcrop. Therefore, we are cautious in our interpretation of streamfunctions near the outcrop and conclude that since streamlines near the wintertime outcrops may not be trustworthy enough to trace mode waters back to their origin, we need an alternative way to identify source regions of mode waters. The method that we adopted is to define the θS relations for individual mode waters in the NPHB climatology and then to compare them with the water properties of the winter mixed layer.

Typical θS relations of mode waters are defined as follows. Since mode water properties tend to change through mixing processes after their formation and also through artificial mixing due to the isopycnal averaging especially across the wintertime outcrops, the least altered portions of mode waters are first defined as the closed low Q cores on the isopycnals. The criteria used to identify the closed low Q areas are listed in Table 1. The θS relationships of these mode water cores in the NPHB climatology are displayed in Fig. 6. Note that, although all the contour maps in the present paper are smoothed with a 2° block mean as described in section 2, we use unsmoothed 1° gridded θ, S, and Q to define these mode-water core properties.

All the 1° grid points that have SST and SSS within the θS ranges of the mode water cores were searched and marked on the maps of the wintertime MLD and SSD (Fig. 7). These grid points satisfy the necessary condition for the mode-water formation site. While all the grid points are not necessarily the formation sites, most of them are consistent with the distribution of the low Q signature of the mode waters. That is, most of these grid points are generally located in the upstream region of the mode water cores although they are not strictly connected to the low Q cores along the streamlines, as we expected. Therefore, we call these grid points the probable formation sites. The ESTMW formation region spans an area from 190° to 210°E along the 25.2 σθ outcrop. The STMW formation area extends from 136° to 142°E along the 25.0 σθ outcrop, from 140° to 160°E along the 25.2 σθ outcrop, and from 145° to 162°E along the 25.4 σθ outcrop. The CMW formation area extends from 143° to 190°E along the 26.0 σθ outcrop, from 146° to 186°E along the 26.2 σθ outcrop, and from 154° to 165°E along the 26.4 σθ outcrop.

We believe that these probable formation sites are in part directly connected to the low Q cores of the mode waters in the real ocean. These actual connections are supported in some cases for STMW and ESTMW by the previous maps of seasonal mean Q climatology such as Suga and Hanawa's (1995) Fig. 8 for the STMW isopycnal and Hautala and Roemmich's (1998) Fig. 9 for the ESTMW isopycnal. There is a hint of the connection between the low Q core of CMW and the probable formation sites in the present Q maps; the relative low Q tongues or bulges extend westward/northwestward from the low Q core to the probable formation sites in Figs. 5h–5k. Some synoptic sections further support this physical connection. WOCE Hydrographic Programme (WHP) P13J section along 165°E in May 1993 captured the Q minimum layer centered on 26.1–26.2 σθ at 36°–37°N, which has the properties of the CMW cores defined in Fig. 6 (Kawabe and Taira 1998). Moreover repeat hydrographic sections along 165°E in May/June from 1996 to 2001 detected CMW at densities between 26.0 σθ and 26.5 σθ and at latitudes between 35° and 40°N every year except 1999 when the section did not cover the latitudes south of 40°N (E. Oka and T. Suga 2003, unpublished manuscript). Since these observations along 165°E were made a couple of months after the CMW formation period, the detected CMW were likely formed near 165°E, consistent with the probable formation sites suggested in Fig. 7.

It should be noted that, since both the isopycnal climatology and the mixed layer climatology are based on data sampled unevenly in time and space, interannual variation in ocean surface/water mass properties likely affects the extent of the probable formation sites inferred above. However, because we cannot tell how it does based on the present climatological analysis, we leave this issue for future study. Analyses of repeat hydrographic sections such as done by E. Oka and T. Suga (2003, unpublished manuscript) will help to clarify this issue.

For comparison, we searched the winter mixed layer climatology derived from WOA94 for grid points with the θS properties of the mode waters defined above (Fig. 8). The number of grid points that have the mode water core properties are considerably lower than those in the NWMLC. While the spatial extent of the probable formation areas of STMW and ESTMW are relatively similar to those in NWMLC, that of CMW is significantly narrower than that found in the NWMLC. The CMW formation sites detected in WOA94 are confined at 180°–188°E along the 26.0 σθ outcrop, at 180°–186°E along the 26.2 σθ outcrop and at 175°–182°E along the 26.4 σθ outcrop except the isolated grid points near the coast of Japan. These probable formation sites are located in the eastward flow with geostrophic speeds of 3–5 cm s−1 and almost directly north of the low Q cores of CMW. Even if there are any unresolved southward flow components, these formation sites can hardly explain all of the low Q cores and cannot explain the relatively low Q tongues or bulges from the core to the west in the NPHB maps. Additionally the CMW formation sites detected in the WOA94 do not explain the recent synoptic observations of CMW at 165°E, while those detected in the NWMLC do. In conclusion, the NWMLC provides a better connection to the NPHB isopycnal climatology than does the WOA94 in the sense that it can better produce all of the mode waters, in particular CMW, required by the NPHB climatology.

It should be noted that the probable formation area of CMW in NWMLC is situated somewhat farther west than that inferred by Suga et al. (1997). They suggested that CMW is formed between 175° and 200°E based on their late winter isothermal layer climatology. However, this formation region is not consistent with the recent synoptic observations as noted by Mecking and Warner (2001) who found that this region may be shifted westward. The present result based on NPHB and NWMLC supports this westward shift.

5. Discussion

Mode waters, represented by low Q cores on isopycnal surfaces and their probable formation sites were presented based on the NPHB isopycnal climatology and NWMLC in the preceding section. The spatial distribution of both the mode waters and their formation sites is inevitably related to the formation mechanisms of the mode waters. Here we qualitatively examine a few aspects of the mode water formation mechanism based on the NWMLC in the framework of subduction theory (e.g., Marshall et al. 1993).

According to the derivation by Williams (1989, 1991), under the assumption of negligible relative vorticity, the potential vorticity of the fluid subducted from the bottom of the steady mixed layer Qm is defined as
i1520-0485-34-1-3-eq1
where f is the planetary vorticity, ρ0 is a reference density, ρm is a mixed layer density, h is a mixed layer thickness, and u and w are the horizontal and vertical velocities at the base of the mixed layer. Subducted fluid acquires a potential vorticity that decreases in value through a decrease in cross-isopycnal flow, −uρm, or through an increase in the subduction rate, −(w + uh), caused by an increase in the vertical pumping term, −w, or in the lateral induction term, −uh (Fig. 9). We examine the probable formation areas identified in the preceding section to see if they are characterized by any of these terms that will lead to low Qm. Note that the annual mean geopotential anomaly at 200 dbar relative to 2000 dbar is superimposed in Fig. 7 as a measure of the flow pattern at the base of the late winter mixed layer.

As demonstrated by Huang and Qiu (1994), the spatial distribution of the vertical pumping term is essentially monotonic in the North Pacific; it shows its highest values at 20°–30°N, 140°–220°E and decreases monotonically towards the north and the east. Therefore, it is reasonable to suppose that a lateral Q minimum is mostly caused by one of the other two terms, that is, the cross-isopycnal flow or the lateral induction.

The probable formation area of ESTMW corresponds with the region of moderate MLD ranging from 100 to 140 m (Fig. 7b). Although MLD shows a slight maximum there, the changes in MLD along streamlines are rather small. On the other hand, as noted in the previous section, SSD gradients are quite small in this area. Figure 7a demonstrates that the density change along the streamlines is indeed small. It is thus suggested that the low Q of ESTMW is caused by the small cross-isopycnal flow, which supports the recent diagnosis of the ocean general circulation model output by Hosoda et al. (2001). Since the low SSD gradients in the ESTMW formation area result from the density-compensating gradients of SST and SSS, the resulting ESTMW tends to have a rather wide range of temperature and salinity on any given isopycnal, as illustrated in Fig. 6.

While the probable formation area of CMW possesses fairly large MLD values ranging from 150 to 225 m, the MLD does not change significantly along the streamlines except for a couple of grid points near 40°N, 160°E (Fig. 7b). These exceptional grid points correspond to the 26.4 σθ outcrop and thus can be sources for the low Q at the 26.4 σθ surface. However, these grid points are too limited in extent in both space and density to produce all of the low Q of CMW. Note, however, that the SSD change along the streamlines is extremely small over the probable formation area of CMW zonally spanning several thousand kilometers at 36°–41°N (Fig. 7a). It is thus inferred that a substantial portion of the low CMW Q results from the small cross-isopycnal flow. Since the general flow is eastward in this region, the zonal band of weak subsurface stratification is likely formed along about 40°N according to a simple subduction scheme assuming small-isopycnal flow. This band of CMW may correspond to the “stability gap” described by Roden (1970) and Yuan and Talley (1996).

The CMW formation process proposed here is contrasted with the situation in several ocean general circulation models (e.g., Kubokawa and Inui 1999; Xie et al. 2000). In those models, the large lateral induction associated with the mixed layer front, or the zone of abrupt change in the mixed layer thickness, is the cause of the low Q in the central part of the basin. This difference mainly comes from the lack of the sharp mixed layer front in the central North Pacific in the NWMLC. Another difference between the CMW in the NWMLC and the equivalent mode water in the models is in the spatial arrangement of the formation sites with respect to their density. The density at the formation sites decreases toward the east in the NWMLC, while it increases toward the east in the models. Note that the former arrangement is consistent with the NPHB isopycnal climatology because the isopycnal low Q signature extends farther to the west with increasing density (see Figs. 5i–k). These differences between the observed climatology and the models indicate an incompleteness in our understanding of CMW formation. Further clarification of the CMW formation process based on observations requires synoptic surveys in the relevant regions during late winter or early spring and/or accumulation of wintertime hydrographic data to produce a better climatology. The model results, on the other hand, have to be compared more closely with the observations to determine whether or not the CMW formation process in the models is realistic.

The probable formation area of STMW is characterized by a very deep mixed layer and associated mixed layer front at its southern flank. Figure 7b demonstrates that the streamlines cross the mixed layer front at 25°–30°N, 135°–170°E. The SSD gradients there are much larger than those in the probable formation areas of ESTMW and CMW. Therefore the low Q of STMW presumably results from a large lateral induction term. The density at the formation sites increases eastward, which resembles what occurs in the models as mentioned above. That is, the formation process of the mode water in the central part of the basin in the models appear to correspond with the STMW formation in the western North Pacific rather than the CMW formation in the central North Pacific.

The above discussion of possible mechanisms of mode water formation supports the good connection between NWMLC and the NPHB isopycnal climatology, even though it is rather qualitative. A more quantitative examination of these mechanisms is beyond the scope of the present study, but is possible by employing calculations of subduction rates and other relevant variables, as done by Huang and Qiu (1994), based on NWMLC and the NPHB climatology. Such calculations are underway and will be reported as a separate work.

Besides the mode-water formation mechanisms, the NPHB isopycnal climatology shows a few notable features associated with the ventilation of the overall subtropical pycnocline. The western region of low Q gradients expands to the east with density increasing from 25.2 to 26.5 σθ as Talley (1988) described for her maps from 25.2 to 26.2 σθ. Talley further noted that much of this low-gradient region that has been identified as the unventilated portion of the wind-driven circulation is actually ventilated through STMW formation. The NPHB climatology indicates that the low-gradient region is also ventilated through CMW formation. Moreover, it reveals that the low Q tongue extends from the lateral Q minimum in the central basin to the tropical western boundary at 26.0–26.5 σθ, as mentioned in section 4a, which implies vigorous ventilation of this region by the advection of CMW.

One of the possible dynamical consequences of this low Q tongue is formation of the subtropical countercurrent in the southern part of the subtropical gyre. Basing their conclusions on an analytical ventilated pycnocline model and an ocean general circulation model, Kubokawa (1999) and Kubokawa and Inui (1999) argue that the stack of the low Q water in the westward flow of the subtropical gyre causes the eastward shallow countercurrent to the south. Aoki et al. (2002), who analyzed several repeat sections and WOCE sections, found a subsurface countercurrent/front along 18°N west of 195°E, immediately to the south of the low Q water including CMW. The present maps show that the low Q tongue of CMW at 26.0–26.3 σθ indeed extends westward along about 20°N, suggesting that CMW is one of agents maintaining the subtropical countercurrent.

It is also notable that Q increases gradually northwestwards from the low Q tongue in the northwestern part of the gyre, suggesting a diffusive flux of the low Q to the northwest or ventilation of the low-gradient region through lateral diffusion. The northwestern part of the gyre at the CMW isopycnals is expected to be the “unventilated” western shadow zone, or the pool region, according to ideal ventilated thermocline theories (Luyten et al. 1983; Talley 1985; Huang and Russell 1994), where a homogenized Q field is predicted (Rhines and Young 1982; Young and Rhines 1982). The dynamical significance of the rather inhomogeneous Q with its systematic lateral gradients in the NPHB climatology will be an interesting subject of future study.

The tongue of CMW extending along the gyre flow contrasts the distribution of CMW as defined by the layer of the vertical temperature gradient minimum, or the thermostad (Suga et al. 1997). The thermostad is spatially confined in the central North Pacific north of 30°N. That is, the CMW defined as the pycnostad and the CMW defined as the thermostad are considerably different from one another, unlike the strong correspondence between STMW defined as a thermostad and as a pycnostad. Authors should be aware of this distinction when describing CMW.

As for the deeper isopycnals, fairly organized patterns appear even below the deepest isopycnal showing the lateral minimum, which contrasts with the situation in the earlier maps. A low Q signature at the 26.6 σθ in the northern subtropical gyre apparently emanates from the northwestern corner of the gyre, off northern Japan, where the isopycnal actually outcrops in late winter according to the NWMLC. Furthermore, it is evident at 26.7–26.8 σθ that the low Q water is derived from the Okhotsk Sea and is advected into the open North Pacific. This feature has been also mentioned by Qu et al. (2001) in discussing their isopycnal maps of the Kuroshio/Oyashio system and is consistent with the generally accepted idea that the Okhotsk Sea provides one of the source waters of North Pacific Intermediate Water (e.g., Talley 1991, 1993; Yasuda 1997). The Q distribution appears fairly uniform at 26.9 σθ and deeper surfaces, which is quite similar to the earlier maps.

In summary, a new winter mixed layer climatology, NWMLC, has been presented, which is naturally connected to the isopycnally averaged climatology of NPHB in the sense that the former is capable of producing all three mode waters appearing in the latter. The probable formation sites of the mode waters have been identified, and the implication of various mode water formation mechanisms has been discussed. It is suggested that the main causes of the low Q of STMW and ESTMW are large lateral induction and small cross-isopycnal flow, respectively. On the other hand, NWMLC suggests that the main cause of the low Q of CMW is small cross-isopycnal flow and not large lateral induction as previously suggested. Last, several features related to the overall subtropical pycnocline ventilation are mentioned, including the fairly organized patterns of the Q distribution in the “unventilated” western shadow zone.

Acknowledgments

We thank Prof. Kimio Hanawa and other members of the Physical Oceanography Group, Tohoku University, for helpful comments and useful discussion. Comments from two anonymous reviewers greatly improved the manuscript. Most of the WOCE data were obtained through the WOCE Hydrographic Programme Office. Dr. M. J. McPhaden kindly provided the CTD data collected by NOAA/PMEL. This study was made as part of Subarctic Gyre Experiment (SAGE), which was financially supported by the former Science Technology Agency and the present Ministry of Education, Culture, Sports, Science, and Technology. This work was supported in part by funds from cooperative program (No. 7, 2001) provided by Ocean Research Institute, University of Tokyo. TS was supported by the Japan Society for Promotion of Science [Grant-in-Aid for Scientific Research (B), No. 13440138]. AMM was supported by NSF Grant OCE-0223421 and NOAA/CICOR Cooperative Agreement NA17RJ1223.

REFERENCES

  • Aoki, Y., , T. Suga, , and K. Hanawa, 2002: Subsurface subtropical fronts of the North Pacific as inherent boundaries in the ventilated thermocline. J. Phys. Oceanogr., 32 , 22992311.

    • Search Google Scholar
    • Export Citation
  • Bingham, F. M., 1992: Formation and spreading of Subtropical Mode Water in the North Pacific. J. Geophys. Res., 97 , 1117711189.

  • Conkright, M. E., and Coauthors, 1999: World Ocean Database 1998 documentation and quality control version 2.0. National Oceanographic Data Center Internal Rep. 14, 113 pp.

    • Search Google Scholar
    • Export Citation
  • Curry, R. G., 1996: HydroBase—A database of hydrographic stations and tools for climatological analysis. Woods Hole Oceanographic Institution Tech. Rep. WHOI-96-01, 44 pp.

    • Search Google Scholar
    • Export Citation
  • Deser, C. M., , A. Alexander, , and M. S. Timlin, 1996: Upper-ocean thermal variations in the North Pacific during 1970–1991. J. Climate, 9 , 18401855.

    • Search Google Scholar
    • Export Citation
  • Hanawa, K., , and I. Hoshino, 1988: Temperature structure and mixed layer over the Izu Ridge. J. Mar. Res., 46 , 683700.

  • Hanawa, K., , and L. D. Talley, 2001: Mode waters. Ocean Circulation and Climate, J. Church et al., Eds., Academic Press, 373–386.

  • Hautala, S. L., , and D. H. Roemmich, 1998: Subtropical mode water in the Northeast Pacific basin. J. Geophys. Res., 103 , 1305513066.

  • Hosoda, S., , S-P. Xie, , K. Takeuchi, , and M. Nonaka, 2001: Eastern North Pacific Subtropical Mode Water in a general circulation model: Formation mechanism and salinity effects. J. Geophys. Res., 106 , 1967119681.

    • Search Google Scholar
    • Export Citation
  • Huang, R. X., , and B. Qiu, 1994: Three-dimensional structure of the wind-driven circulation in the subtropical North Pacific. J. Phys. Oceanogr., 24 , 16081622.

    • Search Google Scholar
    • Export Citation
  • Huang, R. X., , and S. Russell, 1994: Ventilation of the subtropical North Pacific. J. Phys. Oceanogr., 24 , 25892605.

  • Johnson, G. C., , and M. J. McPhaden, 1999: Interior pycnocline flow from the subtropical to the equatorial Pacific Ocean. J. Phys. Oceanogr., 29 , 30733089.

    • Search Google Scholar
    • Export Citation
  • Kara, A. B., , P. A. Rochford, , and H. E. Hurlburt, 2000: An optimal definition for ocean mixed layer depth. J. Geophys. Res., 105 , 1680316821.

    • Search Google Scholar
    • Export Citation
  • Kawabe, M., , and K. Taira, 1998: Water masses and properties at 165°E in the western Pacific. J. Geophys. Res., 103 , 1294112958.

  • Kubokawa, A., 1999: Ventilated thermocline strongly affected by a deep mixed layer: A theory for subtropical countercurrent. J. Phys. Oceanogr., 29 , 13141333.

    • Search Google Scholar
    • Export Citation
  • Kubokawa, A., , and T. Inui, 1999: Subtropical countercurrent in an idealized ocean GCM. J. Phys. Oceanogr., 29 , 13031313.

  • Ladd, C. A., , and L. Thompson, 2000: Formation mechanisms for North Pacific central and eastern subtropical mode waters. J. Phys. Oceanogr., 30 , 868887.

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

  • Levitus, S., , and T. P. Boyer, 1994: Temperature. Vol. 4, World Ocean Atlas 1994, NOAA Atlas NESDIS 4, 111 pp.

  • Levitus, S., , R. Burgett, , and T. P. Boyer, 1994: Salinity. Vol. 3, World Ocean Atlas 1994, NOAA Atlas NESDIS 3, 111 pp.

  • Lozier, S. M., , W. B. Owens, , and R. G. Curry, 1995: The climatology of the North Atlantic. Progress in Oceanography, Vol. 36, Pergamon, 1–44.

    • Search Google Scholar
    • Export Citation
  • Luyten, J. R., , J. Pedlosky, , and H. Stommel, 1983: The ventilated thermocline. J. Phys. Oceanogr., 13 , 292309.

  • Macdonald, A. M., , T. Suga, , and R. G. Curry, 2001: An isopycnally averaged North Pacific climatology. J. Atmos. Oceanic Technol., 18 , 394420.

    • Search Google Scholar
    • Export Citation
  • Marshall, J. C., , A. J. G. Nurser, , and R. G. Williams, 1993: Inferring the subduction rate and period over the North Atlantic. J. Phys. Oceanogr., 23 , 13151329.

    • Search Google Scholar
    • Export Citation
  • McCartney, M. S., 1982: The subtropical recirculation of mode waters. J. Mar. Res., 40 , (Suppl.),. 427464.

  • Mecking, S., , and M. J. Warner, 2001: On the subsurface CFC maxima in the subtropical North Pacific thermocline and their relation to mode waters and oxygen maxima. J. Geophys. Res., 106 , 2217922198.

    • Search Google Scholar
    • Export Citation
  • Nakamura, H., 1996: A pycnostad on the bottom of the ventilated portion in the central subtropical North Pacific: Its distribution and formation. J. Oceanogr., 52 , 171188.

    • Search Google Scholar
    • Export Citation
  • NODC, 1994: World Ocean Atlas 1994 CD-ROM data set documentation. National Oceanographic Data Center Informal Rep. 13, 30 pp.

  • Qu, T., , H. Mitsudera, , and B. Qiu, 2001: A climatological view of the Kuroshio/Oyashio system east of Japan. J. Phys. Oceanogr., 31 , 25752589.

    • Search Google Scholar
    • Export Citation
  • Rhines, P. B., , and W. R. Young, 1982: A theory of the wind-driven circulation. I. Mid-ocean gyre. J. Mar. Res, 40 , (Suppl.),. 559596.

    • Search Google Scholar
    • Export Citation
  • Roden, G. I., 1970: Aspects of the mid-Pacific transition zone. J. Geophys. Res., 75 , 10971109.

  • 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.

    • Search Google Scholar
    • Export Citation
  • Smith, W. H. F., , and P. Wessel, 1990: Gridding with continuous curvature splines in tension. Geophysics, 55 , 293305.

  • Stommel, H. M., 1979: Determination of water mass properties of water pumped down from Ekman layer to the geostrophic flow below. Proc. Natl. Acad. Sci. USA, 76 , 30513055.

    • Search Google Scholar
    • Export Citation
  • Suga, T., , and K. Hanawa, 1990: The mixed layer climatology in the northwestern part of the North Pacific subtropical gyre and the formation area of Subtropical Mode Water. J. Mar. Res., 48 , 543566.

    • Search Google Scholar
    • Export Citation
  • Suga, T., , and K. Hanawa, 1995: The subtropical mode water circulation in the North Pacific. J. Phys. Oceanogr., 25 , 958970.

  • Suga, T., , K. Hanawa, , and Y. Toba, 1989: Subtropical mode water in the 137°E section. J. Phys. Oceanogr., 19 , 16051618.

  • Suga, T., , Y. Takei, , and K. Hanawa, 1997: Thermostad distribution in the North Pacific subtropical gyre: The central mode water and the subtropical mode water. J. Phys. Oceanogr., 27 , 140152.

    • Search Google Scholar
    • Export Citation
  • Suga, T., , A. Kato, , and K. Hanawa, 2000: North Pacific Tropical Water: Its climatology and temporal changes associated with the climate regime shift in the 1970s. Progress in Oceanography, Vol. 47, Pergamon, 223–256.

    • Search Google Scholar
    • Export Citation
  • Talley, L. D., 1985: Ventilation of the subtropical North Pacific: The shallow salinity minimum. J. Phys. Oceanogr., 15 , 633649.

  • Talley, L. D., 1988: Potential vorticity distribution in the North Pacific. J. Phys. Oceanogr., 18 , 89106.

  • Talley, L. D., 1991: An Okhotsk Sea water anomaly: Implications for ventilation in the North Pacific. Deep-Sea Res., 38 , (Suppl.),. S171S190.

    • Search Google Scholar
    • Export Citation
  • Talley, L. D., 1993: Distribution and formation of the North Pacific Intermediate Water. J. Phys. Oceanogr., 23 , 517537.

  • Williams, R. G., 1989: The influence of air–sea interaction on the ventilated thermocline. J. Phys. Oceanogr., 19 , 12551267.

  • Williams, R. G., 1991: The role of the mixed layer in setting the potential vorticity of the main thermocline. J. Phys. Oceanogr., 21 , 18031814.

    • Search Google Scholar
    • Export Citation
  • Xie, S-P., , T. Kunitani, , A. Kubokawa, , M. Nonaka, , and S. Hosoda, 2000: Interdecadal thermocline variability in the North Pacific for 1958–97: A GCM simulation. J. Phys. Oceanogr., 30 , 27982813.

    • Search Google Scholar
    • Export Citation
  • Yasuda, I., 1997: The origin of the North Pacific Intermediate Water. J. Geophys. Res., 102 , 893909.

  • Young, W. R., , and P. B. Rhines, 1982: A theory of wind-driven circulation. II: Gyre with western boundary layers. J. Mar. Res., 40 , 849872.

    • Search Google Scholar
    • Export Citation
  • Yuan, X., , and L. D. Talley, 1992: Shallow salinity minima in the North Pacific. J. Phys. Oceanogr., 22 , 13021316.

  • Yuan, X., , and L. D. Talley, 1996: The subarctic frontal zone in the North Pacific: Characteristics of frontal structure from climatological data and synoptic surveys. J. Geophys. Res., 101 , 1649116508.

    • Search Google Scholar
    • Export Citation
  • Zhang, H-M., , and N. G. Hogg, 1992: Circulation and water mass balance in the Brazil Basin. J. Mar. Res., 50 , 385420.

Fig. 1.
Fig. 1.

The station positions of (a) the dataset used for the annual mean isopycnal climatology and (b) the dataset used for the late winter (Feb–Mar) NWMLC

Citation: Journal of Physical Oceanography 34, 1; 10.1175/1520-0485(2004)034<0003:TNPCOW>2.0.CO;2

Fig. 2.
Fig. 2.

The distribution of (a) the number of stations and (b) the search radius at each grid point for the late winter (Feb–Mar) NWMLC

Citation: Journal of Physical Oceanography 34, 1; 10.1175/1520-0485(2004)034<0003:TNPCOW>2.0.CO;2

Fig. 3.
Fig. 3.

The late winter (Feb–Mar) climatology of the (left) average and (right) standard deviation of (a), (b) mixed layer depth (contour interval of 25 m/10 m with each 50 m/20 m thickened, respectively), (c), (d) sea surface density (contour interval of 0.2 σθ/0.1 σθ with each 1.0 σθ/0.2 σθ thickened, respectively), (e), (f) sea surface temperature (contour interval of 1°C/0.5°C with each 5°C/1°C thickened, respectively), and (g), (h) sea surface salinity (contour interval of 0.1 psu/0.1 psu with each 0.5 psu/0.2 psu thickened, respectively)

Citation: Journal of Physical Oceanography 34, 1; 10.1175/1520-0485(2004)034<0003:TNPCOW>2.0.CO;2

Fig. 3.
Fig. 3.

(Continued)

Citation: Journal of Physical Oceanography 34, 1; 10.1175/1520-0485(2004)034<0003:TNPCOW>2.0.CO;2

Fig. 4.
Fig. 4.

(a) WOA94 mixed layer depth (contour interval of 25 m with each 50 m thickened) in Mar and (b) its difference (contour interval of 20 m with each 40 m thickened) from the NWMLC [present study (P.S.)]. (c) The WOA94 sea surface density (contour interval of 0.2 σθ with each 1.0 σθ thickened) in Mar and (d) its difference (contour interval of 0.1 σθ with each 0.2 σθ thickened) from the NWMLC (P.S.). The dashed contours denote negative differences

Citation: Journal of Physical Oceanography 34, 1; 10.1175/1520-0485(2004)034<0003:TNPCOW>2.0.CO;2

Fig. 5.
Fig. 5.

The annual mean climatology of Q (10−12 m−1 s−1) on the isopycnal surfaces (a) 24.2 σθ, (b) 24.6 σθ, (c) 25.0 σθ, (d) 25.2 σθ, (e) 25.4 σθ, (f) 25.6 σθ, (g) 25.8 σθ, (h) 26.0 σθ, (i) 26.2 σθ, (j) 26.4 σθ, (k) 26.5 σθ, (l) 26.6 σθ, (m) 26.7 σθ, (n) 26.8 σθ, (o) 26.9 σθ, and (p) 27.0 σθ. The contour interval is 100 × 10−12 m−1 s−1 with each 200 × 10−12 m−1 s−1 thickened. Additional dashed contours at intervals of 25 × 10−12 m−1 s−1 are drawn for Q lower than 300 × 10−12 m−1 s−1. Thick white contours denote the winter outcrop of each isopycnal. Thin white contours represent pressure anomaly streamfunction (contour interval of 1 m2 s−2) relative to 2000 dbar. The region where the isopycnal is defined is shaded. The darkest, the second darkest, and the third darkest shades denote the regions with Q lower than 200 × 10−12 m−1 s−1, from 200 × 10−12 to 300 × 10−12 m−1 s−1, and from 300 × 10−12 to 500 × 10−12 m−1 s−1, respectively

Citation: Journal of Physical Oceanography 34, 1; 10.1175/1520-0485(2004)034<0003:TNPCOW>2.0.CO;2

Fig. 5.
Fig. 5.

(Continued)

Citation: Journal of Physical Oceanography 34, 1; 10.1175/1520-0485(2004)034<0003:TNPCOW>2.0.CO;2

Fig. 5.
Fig. 5.

(Continued)

Citation: Journal of Physical Oceanography 34, 1; 10.1175/1520-0485(2004)034<0003:TNPCOW>2.0.CO;2

Fig. 5.
Fig. 5.

(Continued)

Citation: Journal of Physical Oceanography 34, 1; 10.1175/1520-0485(2004)034<0003:TNPCOW>2.0.CO;2

Fig. 6.
Fig. 6.

Potential temperature–salinity diagram for the low-Q water representing mode water cores. Circles, squares, and triangles represent the STMW, CMW, and ESTMW cores, respectively

Citation: Journal of Physical Oceanography 34, 1; 10.1175/1520-0485(2004)034<0003:TNPCOW>2.0.CO;2

Fig. 7.
Fig. 7.

Grid points in the NWMLC, which have properties corresponding to those of the mode water cores as defined in Fig. 6, superimposed on the maps of (a) the sea surface density (contour interval of 0.2 σθ with each 1.0 σθ thickened) and (b) the mixed layer depth (contour interval of 25 m with each 100 m thickened) in the late winter. Circles, squares, and triangles represent the STMW, CMW, and ESTMW cores, respectively. The annual mean climatology of the geopotential anomaly (contour interval of 1 m2 s−2) at 200 dbar relative to 2000 dbar is also superimposed on both maps

Citation: Journal of Physical Oceanography 34, 1; 10.1175/1520-0485(2004)034<0003:TNPCOW>2.0.CO;2

Fig. 8.
Fig. 8.

Grid points in the WOA94 Mar mixed layer climatology, which have properties corresponding with those of the mode water cores as defined in Fig. 6, superimposed on the maps of (a) the sea surface density and (b) the mixed layer depth in Mar. Circles, squares, and triangles represent the STMW, CMW, and ESTMW cores, respectively

Citation: Journal of Physical Oceanography 34, 1; 10.1175/1520-0485(2004)034<0003:TNPCOW>2.0.CO;2

Fig. 9.
Fig. 9.

Schematics showing the terms determining Q of the fluid subducted from the bottom of steady mixed layer: cross-isopycnal flow, −uρm; lateral induction, −uh; and vertical pumping, −w

Citation: Journal of Physical Oceanography 34, 1; 10.1175/1520-0485(2004)034<0003:TNPCOW>2.0.CO;2

Table 1.

Definition of the mode water cores

Table 1.
Save