Subduction of the North Pacific Mode Waters in a Global High-Resolution GCM

Tangdong Qu International Pacific Research Center, SOEST, University of Hawaii, Honolulu, Hawaii

Search for other papers by Tangdong Qu in
Current site
Google Scholar
PubMed
Close
,
Shang-Ping Xie International Pacific Research Center, SOEST, University of Hawaii, Honolulu, Hawaii

Search for other papers by Shang-Ping Xie in
Current site
Google Scholar
PubMed
Close
,
Humio Mitsudera Frontier Research System for Global Change, Tokyo, Japan, and International Pacific Research Center, SOEST, University of Hawaii, Honolulu, Hawaii

Search for other papers by Humio Mitsudera in
Current site
Google Scholar
PubMed
Close
, and
Akio Ishida Frontier Research System for Global Change, Tokyo, and Japan Marine Science and Technology Center, Yokosuka, Japan

Search for other papers by Akio Ishida in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

The annual subduction rate in the North Pacific is estimated using five-day outputs from a high-resolution general circulation model (GCM). Two maxima (>200 m yr−1) are found in the western North Pacific: one is responsible for the formation of the subtropical mode water (STMW) and the other for the formation of the central mode water (CMW). A local maximum (>75 m yr−1) is also found in the formation region of the eastern subtropical mode water (ESMW). These results are compared with a calculation using the winter mixed layer depth and annual mean velocity fields to examine the effect of mesoscale eddies. Although the mesoscale eddies do not markedly affect the general subduction pattern, they enhance the annual subduction rate by up to 100 m yr−1 in the formation region of the STMW/CMW, a 34% increase in a regional average (30°–44°N, 140°E–170°W). Further analysis shows that the effects of the mean seasonal cycle and smaller-scale (<30 days) eddies are generally small. The authors suggest that the two peaks in the subduction rate are related to a double-front structure on the intergyre boundary in the western North Pacific.

Corresponding author address: Dr. Tangdong Qu, IPRC-SOEST, University of Hawaii at Manoa, 2525 Correa Road, Honolulu, HI 96822. Email: tangdong@soest.hawaii.edu

Abstract

The annual subduction rate in the North Pacific is estimated using five-day outputs from a high-resolution general circulation model (GCM). Two maxima (>200 m yr−1) are found in the western North Pacific: one is responsible for the formation of the subtropical mode water (STMW) and the other for the formation of the central mode water (CMW). A local maximum (>75 m yr−1) is also found in the formation region of the eastern subtropical mode water (ESMW). These results are compared with a calculation using the winter mixed layer depth and annual mean velocity fields to examine the effect of mesoscale eddies. Although the mesoscale eddies do not markedly affect the general subduction pattern, they enhance the annual subduction rate by up to 100 m yr−1 in the formation region of the STMW/CMW, a 34% increase in a regional average (30°–44°N, 140°E–170°W). Further analysis shows that the effects of the mean seasonal cycle and smaller-scale (<30 days) eddies are generally small. The authors suggest that the two peaks in the subduction rate are related to a double-front structure on the intergyre boundary in the western North Pacific.

Corresponding author address: Dr. Tangdong Qu, IPRC-SOEST, University of Hawaii at Manoa, 2525 Correa Road, Honolulu, HI 96822. Email: tangdong@soest.hawaii.edu

1. Introduction

Layers of weak stratification and homogenous properties called mode waters are often observed in the thermocline of the oceans [see a review by Hanawa and Talley (2001)]. Mode waters have recently been implicated as an important agent for subsurface decadal variability (Yasuda and Hanawa 1997; Xie et al. 2000). Key to the formation of the mode waters is the volume flux of mixed layer water entering the thermocline per unit horizontal area, or the so-called subduction rate. Theoretical studies of the subduction rate date back to Iselin (1939), who first suggested that water is pushed down through the surface layer by Ekman pumping. Recent studies, however, indicate that the amount of water actually subducted into the thermocline is considerably larger than can be accounted for by Ekman pumping alone. Stommel (1979) recognized that it is actually the flow leaving the base of a deep winter mixed layer that ventilates the thermocline (the “mixed layer demon” hypothesis). This hypothesis has been followed by a number of ideal-fluid pycnocline models that diagnose the subduction rate (e.g., Woods 1985; Williams 1991; Williams et al. 1995; Marshall 1997). All of these models suggest that the volume of subducted fluid within the subtropical gyre is greatly enhanced by lateral induction, or horizontal advection across the sloping mixed layer by the gyre circulation. In most mode-water formation regions, the subduction rate due to lateral induction is much greater than that resulting from Ekman pumping (see Hanawa and Talley 2001).

Estimates of the subduction rate and, in particular, its component due to lateral induction, require detailed information on the mixed layer depth and velocity fields. Observational data are generally too sparse to provide such information. Based on a simplified version of the model proposed by Woods (1985; see also Williams 1991), Qiu and Huang (1995) estimated the subduction rate using the Levitus (1982) climatology and Hellerman and Rosenstein (1983) wind stress data. According to their calculation, the maximum subduction rate is O(100 m yr−1), and the total volume subducted from the mixed layer into the permanent thermocline is 35.2 Sv (Sv ≡ 106 m3 s−1) over the entire North Pacific. The subduction rate is biased toward a density range between 25.0 and 25.2 σθ where the subtropical mode water (STMW: Masuzawa 1969) is found. In the density range of the central mode water (CMW: Nakamura 1996; Suga et al. 1997; Talley 1999), Qiu and Huang's (1995) subduction rate (their Fig. 13) does not show a peak. This result differs from what one would expect based on the mode water distribution (Talley 1999). The lack of a subduction peak at the CMW density could be due to poor quality climatological data in the subpolar region. Since the objective analysis in preparing such products often smoothes the data over a horizontal distance as large as 700 km, certain features derived from such analyzed datasets may be questionable.

Another important issue in estimating the subduction rate is the effect of high-frequency variability, which Qiu and Huang (1995) did not address. Mode-water formation regions see large seasonal variations and are populated with mesoscale eddies. As a result, the subduction rate derived based on the mean flow may not be representative (Marshall 1997). The best studied region in this regard is the North Atlantic, where a large eddy contribution to subduction has been identified (Williams 1989; Williams et al. 1995). In a recent model simulation with idealized forcing and geometry, Hazeleger and Drijfhout (2000) suggested that eddies can enhance the annual subduction rate by almost a factor of two south of the Gulf Stream extension. To the best of our knowledge, no similar study has been carried out for the North Pacific.

In this study, we analyze the results from a high-resolution general circulation model (GCM) to seek a better understanding of subduction processes, in particular, the effect of eddies on the formation of the North Pacific mode waters. The rest of the paper is organized as follows. After a brief model description in section 2, we present some basic characteristics of the model solution in section 3. Section 4 determines the subduction rate, and section 5 further partitions the annual subduction into several density groups. Potential vorticity distributions indicating the formation of mode waters on a set of density surfaces are presented in section 6. The effect of mesoscale eddies on subduction in the formation region of the STMW/CMW is discussed in section 7. The results are summarized in section 8.

2. Model and data processing

Here we briefly describe the model used in this study. See Ishida et al. (1998) for more details. The model is developed at Japan Marine Science and Technology Center, based on Modular Ocean Model version 2 (MOM2). It has a horizontal resolution of ¼ degree both in latitude and longitude, and has 55 levels in the vertical whose spacing varies from 10 m near the sea surface to about 50 m in the intermediate layers and about 400 m at 6000 m depth. The model covers the whole World Ocean except for the Arctic Ocean, with its northern and southern boundaries closed at 75°N and 75°S, respectively. It has a realistic coastline and bottom topography derived from the National Geophysical Data Center dataset (ETOPO5) with a 5′ × 5′ resolution.

The model is spun up from the initial state at rest with Levitus (1982) annual mean temperature and salinity. In the first two model years, the model is forced by the Hellerman and Rosenstein (1983) annual mean wind stress, and the heat and freshwater flux is implemented as a linear restoring of temperature and salinity in the first model level toward the Levitus annual mean climatoloy. Harmonic viscosity is used during this period to avoid initial numerical instability from the rest state. Then, the model is forced by the Hellerman–Rosenstein monthly wind stress for 18 years. At the sea surface, temperature and salinity are restored toward Levitus monthly climatology. The restoring coefficient is chosen such that the e-folding time is 30 days for a mixed layer of 50 m deep. For horizontal turbulent mixing, a highly scale-selective biharmonic operator is used, with a coefficient of −1 × 1019 cm4 s−1 for both momentum and tracers. The vertical mixing follows the Pacanowski and Philander (1981) formulation, and no isopycnal thickness mixing parameterization is incorporated. There is no significant drift in the model after the 10-yr integration.

For this study results from the last model year are used, which include instantaneous velocity, temperature, and salinity every five days. At a single station/grid point without information on its surroundings, data should cover at least several eddy life cycles to produce a statistically reliable mean field. Given that the mesoscale eddies in the western North Pacific have a typical timescale of 100–200 days (e.g., Ebuchi and Hanawa 2000; Feng et al. 2000), the one year of data used in this study is not long enough for this purpose. Here we take an alternative approach taking advantage of the fact that mesoscale eddies are quasiperiodic in the zonal direction. We calculate the decorrelation scale of eddies, defined as the e-folding distance of the velocity correlation coefficient at 100 m. The decorrelation scale is generally zonally oriented, and has a maximum (∼1000 km) at the equator (not shown). In the formation region of the North Pacific mode waters (20°–50°N), the typical decorrelation scale (regional average) is 1.5° in latitude and 3° in longitude. We define the mean as the smoothed field using a Gaussian filter with an e-folding scale of 3° in the zonal direction; no smoothing is applied in the meridional direction to preserve the sharp Kuroshio Extension and Oyashio Front. This approach reduces the effect of eddies and thus results in a better statistical mean than that without smoothing as will be seen in section 7.

3. Basic characteristics of the model solution

a. Mean kinetic energy

The mean kinetic energy (MKE) at 5 m is presented to show the basin-scale circulation in the North Pacific (Fig. 1a). Here, MKE is estimated from the annual mean velocity field, namely, MKE ≡ u2 + υ2, where u and υ denote the zonal and meridional components of mean velocity, respectively. The MKE field indicates that the model has a reasonable representation of the South Equatorial Current on the equator, the North Equatorial Current at 10°–20°N, the Mindanao Current and its eastward turning in the Tropics, and the northward-flowing Kuroshio along the western boundary, with maximum MKE values exceeding 750 cm2 s−2 in these regions.

The Kuroshio leaves the coast of Japan at about 34°N, in reasonable agreement with observations, and flows eastward as a free jet, the Kuroshio Extension, to as far as the international date line. The Oyashio leaves the coast of Japan near 40°N. Though this latitude is slightly farther northward than observations (Qu et al. 2001), the Oyashio Front extending eastward along 42°N is a reasonable representation of the true ocean. The separation point of the Kuroshio/Oyashio is a common problem for many GCMs (e.g., Semtner and Chervin 1992), and the success of the present model in this regard is encouraging and gives us some confidence in the fidelity of the model simulation.

b. Eddy kinetic energy

The quarter-degree horizontal resolution allows the model to represent eddies of various spatial scales. Here we define eddies as perturbations from the annual mean. Eddies are very active near the western boundary east of Japan, along the Kuroshio Extension at 33°N and the Oyashio Front at 43°N (Fig. 1b), where eddy kinetic energy (EKE) is of the order 1500 cm2 s−2 due to the vacillation of the Kuroshio Extension, fluctuation of the recirculation gyre, and the presence of warm and cold core rings. The timescales of these fluctuations are typically several months to annual. The EKE is also high near the equator. In the eastern equatorial Pacific, this is due to tropical instability waves with a maximum amplitude around 3°N, while in the west, large variability in the western boundary currents is responsible for high EKE. Most of these features compare favorably with those from altimeter data (Ishida et al. 1998).

c. Winter mixed layer depth

Mixed layer depth (MLD), defined as the depth at which water density is 0.1 σθ heavier than at the sea surface, reaches its seasonal maximum in March (Fig. 2). March MLD is generally shallow (about 50 m) along the southern rim of the subtropical gyre, but becomes increasingly deeper toward higher latitudes. In the mixed water region east of Japan, between the Kuroshio Extension and Oyashio Front, the March MLD is of the order 200 m, with its maximum exceeding 300 m on the warmer side of the Oyashio Front. Separating this deep mixed layer region (27°–42°N, west of 170°W) from the rest of the North Pacific is a narrow transition zone called the MLD front (Xie et al. 2000), a key to the formation of mode waters (Kubokawa 1999). In the southeastern basin, there is a local maximum (>75 m) centered at about 30°N, 140°W, which is partly responsible for the formation of the eastern subtropical mode water (ESMW: Talley 1988; Hautala and Roemmich 1998; Hosoda et al. 2001).

For comparison, we have also included the March MLD calculated from the World Ocean Atlas 1998 (WOA98) using the same density criterion (Fig. 2b). The largest discrepancy is seen in the northwestern subtropical gyre, where the model overestimates the depth by about 100 m but underestimates the density of the mixed layer by about 0.5 σθ. The warming bias in the western subtropical gyre is a deficiency common to all GCMs that restore SST back to the observed climatology (e.g., Xie et al. 2000), and may enhance the winter convection, leading to a deeper winter MLD.

Discrepancy in MLD is also apparent in the subpolar region, where model MLD (<50 m) is shallower than observations by at least a factor of 2. This discrepancy could be a result of the vertical mixing parameterization (Pacanowski and Philander 1981) whose validity at high latitudes has not been tested. It must be noted, however, that the MLD field (Fig. 2b) derived from the highly smoothed climatological data may also contain large errors. Preliminary examination of the raw data for the WOA98 indicates that the data coverage is rather sparse in the subpolar region, especially in winter (Qu et al. 2001). As a result of smoothing over more than 700 km, the MLD front along the northern rim of the subtropical gyre might have been significantly weakened.

As in observations, the model MLD exhibits a large annual excursion (not shown) in response to the surface buoyancy forcing. It starts to shoal in mid and late spring (Apr–May), and deepens after reaching its minimum in summer (Jul–Aug). The rapid shoaling of MLD allows winter mixed layer water to be subducted into the thermocline and carried away by the subtropical gyre (discussed in section 4).

d. Thermocline structure

Sections of potential temperature, salinity, and density from the smoothed annual mean fields are presented to show the thermocline structure associated with the Kuroshio Extension and Oyashio Front at 150°E (Fig. 3). The overall structure in the model compares reasonably well with observations (Qu et al. 2001). Throughout the section water temperature is dominated by an upward slope of isotherms toward the north, indicative of a broad eastward flow that forms the boundary between the subtropical and subpolar gyres. In the south the seasonal thermocline overrides a thick layer of weak stratification and high salinity indicative of the STMW. The North Pacific Intermediate Water (NPIW: Talley 1999) stands out as a salinity minimum extending southward beneath the ventilated thermocline.

On a closer inspection of density field, we see a double-front structure in the main thermo/pycnocline. One front resides at 34°N corresponding to the Kuroshio Extension, and the other lies at 43°N corresponding to the Oyashio Front. This double-front structure is a well-established feature that characterizes the intergyre boundary of the North Pacific circulation (e.g., Joyce 1987; Schmitz et al. 1987; Qu et al. 2001) and, as will be shown below, is an important factor for the subduction of the North Pacific mode waters. In the model, the depth of the main thermo/pycnocline show steplike jumps at these fronts that are much stronger than the long-term mean observations. Large interannual variability has been observed in these fronts (e.g., Qiu and Miao 2000), which may have weakened these fronts in the climatological average.

4. Subduction rate

By definition, the instant subduction rate S(t) or the volume flux of mixed layer water entering the thermocline per unit horizontal area is the sum of the vertical velocity of water parcel at the base of the mixed layer, −wmb; the rate of mixed layer shoaling −∂hm/∂t, and the rate at which the water is swept beneath the shoaling mixed layer base by horizontal currents −umb · ∇hm (Woods 1985; Cushman-Roisin 1987; Williams 1991; Qiu and Huang 1995; Williams et al. 1995),
i1520-0485-32-3-746-e1
where the subscript mb denotes the mixed layer base. The annual subduction rate Sann can be obtained by integrating the instant subduction rate over one year (T) from the end of the first winter t1 to that of the second winter t2 in a Lagrangian framework. That is,
i1520-0485-32-3-746-e2
where the first term on the right-hand side (rhs) represents the contribution from vertical pumping at the base of the mixed layer, and the second term the contribution from lateral induction due to the sloping of the mixed layer base. In the following, we trace water parcels released at the base of the March mixed layer using various velocity fields to examine the contributions to the annual subduction rate of the mean flow, seasonal cycle, and meso- and small-scale eddies.

a. Subduction rate by the mean flow

Equation (2) is a straightforward calculation but requires detailed information on the time–space variations of the mixed layer depth and velocity fields, which is usually not available from observational data. Qiu and Huang (1995) suggested a simplified version of Eq. (2) to circumvent this problem by advecting the Lagrangian particles with the annual mean velocity field. For the first term on the rhs of Eq. (2), they substituted the annual mean velocity wann at the bottom of March mixed layer for wmb and ignored the changes of mixed layer depth.

Following Qiu and Huang (1995), we estimate the annual subduction rate using March mixed layer depth and annual mean velocity fields (hereinafter called the annual subduction rate by the mean flow and denoted as Smean). Here we trace water parcels along Lagrangian trajectories with a time step of 5 days. The annual subduction rate due to Ekman pumping varies from 0 to 50 m yr−1 (Fig. 4a) in the subtropical gyre. Lateral induction exceeds 100 m yr−1 (Fig. 4b) in the northwestern subtropical gyre, as a result of the strong horizontal gradient of the winter MLD (Fig. 2a). The lateral induction rate decreases rapidly toward the southeast and becomes less important than the contribution of the vertical pumping in the southeastern subtropical gyre.

The spatial distribution of annual subduction rate (Fig. 4c) shows several features resembling those derived from climatological data (Qiu and Huang 1995, their Fig. 11c). In both the model and observations, there is a local maximum at about 30°N between 150° and 170°E that exceeds 100 m yr−1. Farther to the east, another local maximum is seen at 180°–160°W; its magnitude in the model (>200 m yr−1) is four times greater than that in Qiu and Huang's climatological map (>50 m yr−1) however. This subduction rate is also high (>200 m yr−1) around 42°N in the model. In contrast to a weak local maximum (>25 m yr−1) in Qiu and Huang's climatological map, the model produces a strong and extensive subduction zone on the warmer side of the subpolar front.

A tongue of negative subduction rate extends eastward from the coast of Honshu to about 170°E along 35°N, separating the mode-water formation region into two parts. The southern part, including the first two maxima in the subduction rate cited above, has a winter surface density ranging from about 24.4 to 25.4 σθ (Fig. 2b). The underlying geostrophic flow in this region has a strong southward component associated with the recirculation of the western subtropical gyre (Fig. 5). Judged from its geographical location and dynamical characteristics, this part of the subduction corresponds to the formation of the STMW (see Fig. 2 of Talley 1999). The northern subduction maximum is located on the warmer side of the subpolar front, covering a range of winter surface density between 26.0 and 26.4 σθ. Water in this density range corresponds to the CMW and is advected by the eastward North Pacific Current (Fig. 5; Suga et al. 1997; Talley 1999).

Between 35° and 40°N, the MLD decreases rapidly eastward around 170°W (Fig. 2a), but the subduction rate between 25.4 and 26.0 σθ is relatively small. Because the local vertical pumping is weak, the annual subduction rate is primarily determined by the MLD difference between the first and second winters on the water parcel trajectory [i.e., the second term on the rhs of Eq. (2)]. This MLD difference, however, depends not only on the strength of the MLD front but also on the distance that a water parcel moves during the year. Between 35° and 40°N there is a zone of weak eastward flow, and the flow even reverses its direction toward the west in some cases (Qu et al. 2001). As a result, the distance that a water parcel moves from its release point is short (Fig. 5). This short advection distance leads to a low subduction rate despite the presence of a MLD front around 170°W.

In the southeastern North Pacific, a local maximum (>75 m yr−1) is centered around 25°N, 140°W and falls in the density range (∼25.0 σθ) of the ESMW. In addition to this local subduction maximum, the wide spacing between the 24.5 and 25.5 isolines of winter surface density (Fig. 2a) also helps the formation of the ESMW (Xie et al. 2000; Ladd and Thompson 2001; Hosoda et al. 2001). In Fig. 4c, we see that about two-thirds of the annual subduction in the ESMW formation region is due to vertical pumping, in contrast with the formation of the STMW/CMW that is dominated by lateral induction. On close inspection, however, we find that the ESMW formation region lies actually in a much broader downwelling region of a vertical pumping rate greater than 50 m yr−1 that stretches from 15° to 35°N. It is the lateral induction that localizes the net subduction maximum around the ESMW formation by enhancing (weakening) the vertical pumping contribution there (elsewhere).

b. Total subduction rate

In the region where the North Pacific mode waters are formed, the ocean is populated with variabilities of various timescales (Fig. 1b), and as a result, the annual subduction rate described above may not be representative. Using instantaneous model outputs, we are able to calculate the total subduction rate (denoted as Stot below) without approximations in Eq. (2) and include the contribution from variabilities. As in the previous subsection, water parcels are first released at the base of the mixed layer in March, and then traced for one year with a time step of 5 days. The total subduction rate (Fig. 6) shows essentially the same pattern as that computed based on the mean flow (Fig. 4), except that the two maxima along 30°N in Fig. 4 show a tendency to merge together in Fig. 6. The maximum subduction rate now exceeds 250 m yr−1 in the STMW/CMW formation region, and the tongue of negative values around 35°N is narrowed.

c. Eddy-induced subduction rate

The total subduction rate, Stot, described above consists of contributions from the mean flow and variabilities. In the following, we refer to the contribution due to variabilities as the eddy-induced subduction rate, measured simply as the difference between Stot and Smean. The largest eddy contribution is found in the formation region of the STMW/CMW, with its magnitude exceeding 100 m yr−1 (Fig. 7a). The eddy contribution drops to less than 25 m yr−1 in other parts of the subtropical gyre, where positive and negative values tend to cancel each other and the net eddy effect on large-scale subduction is weak.

Eddies affect the annual subduction rate in two ways: by modifying the MLD, and hence the vertical pumping at the base of mixed layer [the first term on the rhs of Eq. (2)], and by modifying the velocity field that advects water parcels [the second term on the rhs of Eq. (2)]. To estimate the relative importance of these two processes, we recalculate the subduction rate using instantaneous MLD and annual mean velocity fields in one case and using the March MLD and instantaneous velocity fields in another. The differences of these estimates from Smean are regarded as the contributions from MLD fluctuation (Fig. 7b) and velocity fluctuation (Fig. 7c), respectively. The annual excursion of MLD has little effect on the annual subduction rate (Fig. 7b). The maximum contribution (>20 m yr−1) of this process is seen on the warmer side of the Kuroshio Extension, but not at the Oyashio Front where Ekman pumping is close to zero during most seasons of the year. Instead, at the Oyashio Front we see the total eddy-induced subduction rate (Fig. 7a) dominated by velocity fluctuation (Fig. 7c) through lateral induction.

The contribution from velocity fluctuations can be further decomposed into mean seasonal cycle, mesoscale eddy, and small-scale eddy components. Here the mean seasonal cycle is defined as the monthly mean velocity field; mesoscale eddies as variabilities less than 600 km in the zonal direction and more than 30 days in period; small-scale eddies as variabilities with periods shorter than a month. We calculate the subduction rate in two ways: using the (zonally smoothed) monthly mean velocity field (denoted as Smon) and using the zonally unsmoothed monthly mean velocity field (denoted as Sunmon). The mean seasonal cycle, Smo = SmonSmean, contributes little to the annual subduction rate (Fig. 8a). The mesoscale eddy contribution (Fig. 8b), Sme = SunmonSmon, accounts for most of the eddy-induced subduction rate (Fig. 7a). The small-scale eddy contribution, Sse = StotSunmon, is randomly distributed and has little effect on the large-scale subduction pattern (Fig. 8c).

5. Annual subduction

By integrating the annual subduction rate over the whole North Pacific in 10°–50°N, we obtain an annual subduction volume of 61.6 Sv, about 75% larger than that (35.2 Sv) derived from the climatological data (Qiu and Huang 1995). To this number, the annual mean flow contributes 49.0 Sv, mean seasonal cycle 0.2 Sv, mesoscale eddies 7.8 Sv, and small-scale eddies 4.6 Sv (Table 1).

Using the winter surface density field (Fig. 2a), the annual subduction can be further partitioned among density groups. Two peaks stand out in subduction volume sorted by winter density (Fig. 9): one corresponds to the production of the STMW between 24.4 and 25.4 σθ, and the other has a density range of 26.0–26.4 σθ, corresponding to the CMW in the model. The subduction in the STMW density range is 30.4 Sv, which is considerably larger than that (8.4 Sv) associated with the CMW. Most of the eddy-induced subduction occurs in the density range of 24.4–26.4 σθ (shaded bars in Fig. 9). For a regional average over the formation region of the STMW/CMW (30°–44°N, 40°E–170°W), the eddy contribution (12.4 Sv) accounts for 43% of the total subduction (28.9 Sv) and most of this contribution (9.7 Sv) is due to mesoscale eddies.

6. Mode water formation

The potential vorticity (PV), Q, of subducted water is related to the subduction rate by (Williams 1989, 1991)
i1520-0485-32-3-746-e3
where f is the planetary vorticity, ρ is a reference density, and ρm is the density of mixed layer water. In other words, the potential vorticity of subducted water tends to decrease with increasing S(t), either through enhanced vertical pumping or lateral induction. Since strong subduction is associated with lateral induction across large winter MLD gradient, mode waters with small PV tend to form where the winter MLD front intersects with the outcrop line (Kubokawa 1999; Xie et al. 2000).

Figure 10 shows the March potential vorticity on six isopycnals between 24.7 and 26.2 σθ. The 200 m MLD contour is plotted to show the MLD front. A distinct PV minimum is found on each of these isopycnals at the intersection between the MLD front and outcrop line. As noted by Masuzawa (1969), mode waters acquire progressively higher densities as their formation regions move to the east. These mode waters are distributed far beyond their formation regions as a result of the advection by geostrophic flow. On the light isopycnals, the western part of the low-PV (<2.0 × 10−10 m−1 s−1) region is associated with the Kuroshio Extension recirculation gyre, and the STMW is well defined on density surfaces around 24.7 σθ, somewhat lighter than observations (Hanawa 1987; Hanawa and Talley 2001) due to the warm bias of the model mixed layer temperature (section 3c). On the deeper isopycnals (σθ > 25.6), the low-PV tongue extends eastward before turning southwestward. The broadest spatial extent of low-PV water occurs at 26.2 σθ, corresponding to the CMW that marks the bottom of the ventilated thermocline.

The formation of low-PV (<2.0 × 10−10 m−1 s−1) water continues until early May when MLD starts to shoal rapidly (not shown). After this time, the outcrop line moves to the north and the mixed layer front shrinks in response to the surface heating. As a result, water entering the thermocline in winter becomes completely shielded from the mixed layer and its variations.

7. Discussion

To have a better understanding of how eddies modify the annual subduction rate, we show below the velocities along two zonal sections within the formation region of the STMW/CMW: one is along 32°N and the other along 42°N. In both of these sections the eddy-induced subduction rate exceeds 100 m yr−1 (Fig. 7a).

In the formation region of the STMW at 32°N, the (zonally smoothed) annual mean velocity consists of a swift and narrow northward flow corresponding to the Kuroshio near the western boundary and a weak and broad southward flow representing the Kuroshio Extension recirculation gyre farther to the east (Fig. 11a). The unsmoothed annual mean velocity, by contrast, is somewhat noisy, with northward and southward flows alternating throughout the section (Fig. 11b). This difference clearly demonstrates the influence of mesoscale eddies with a typical spatial scale of 300–600 km. For each individual month, the smoothed velocity shows essentially the same pattern as that shown in Fig. 11a, while the structure of the unsmoothed field differs one month from another due to the propagation of mesoscale eddies (not shown). The upper-layer (0–500 m) transport through the section varies from about 43 Sv in July to 53 Sv in January (Fig. 12a).

This seasonal cycle of transport within the Kuroshio Extension recirculation gyre has not been previously reported. To gain a general idea of how representative this model result is, we also calculate the Ekman transport from Hellerman and Rosenstein (1983) wind stress and the geostrophic flow in the upper 500 m from WOA98 by assuming a 2000-db reference level. The result (geostrophic + Ekman) shows a remarkable agreement with that derived from the model (Fig. 12b). In both cases, the flow across the section has a larger southward transport in winter than in summer. In a quantitative sense, however, we notice that the model transport is greater by about 20 Sv than the observed transport. This discrepancy could be interpreted as a result of meridional smoothing used in preparing the climatological dataset. Indeed, if a Gaussian filter with e-folding scale of 1.5° latitude were further applied to the smoothed velocity field in the model, the two transports would become closer. Another possible contribution to this discrepancy is the exclusion of the deep component of the geostrophic calculation.

In the formation region of the CMW at 42°N, the smoothed annual mean velocity is considerably weaker (Fig. 13a), and the spatial scale of mesoscale eddies is somewhat smaller (Fig. 13b). The northward transport (0–500 m) through this section has a maximum (about 8 Sv) in summer and a minimum (close to zero) in winter (Fig. 14), yielding a mass transfer of about 4 Sv from the surface mixed layer into the thermocline on the annual average.

The seasonal variation of transport in the formation region of the STMW/CME seems to be dominated by the surface Ekman current. Along both sections, we see that the geostrophic transport does not change significantly throughout the year (Figs. 12b and 14b), and this to a large extent explains why the inclusion of mean seasonal cycle in the velocity field has little impact on the annual subduction rate of the region (Table 1).

The effect of mesoscale eddies has been studied by several earlier studies (Williams et al. 1995; Marshall 1997; Hazeleger and Drijfhout 2000), though none of these studies uses results from high-resolution, realistic GCMs. Among others, Marshall suggests that, in regions of intense baroclinic instability and strong MLD front, an ensemble of mesoscale eddies may provide an additional advection by the “bolus velocity.” This additional advection may enhance the annual subduction rate by a value that is potentially of the same order and cannot be neglected.

Here we emphasize that the subduction of mode waters is a one-way process; the effect of the southward flow of a mesoscale eddy that transports a particular water parcel in a Lagrangian sense is not counterbalanced by the effect of the northward flow of the eddy. In the formation region of the STMW, for example, the southward flow of a mesoscale eddy can force the subducted water to move as far as several hundred kilometers south of the MLD front. From there, this water is advected farther southwestward into the thermocline by geostrophic flow in the recirculation gyre. Indeed, we see that with the smoothed monthly velocity field (Fig. 15a), eastward trajectories are prominent along the Kuroshio Extension (see also Fig. 5), whereas southward trajectories are prominent with the unsmoothed monthly velocity field (Fig. 15b). Although the northward flow component of mesoscale eddies may bring some water back, much of it has been far away from its formation region and cannot be reentrained into the shallow mixed layer at its new location in the following winter. As such, the mesoscale eddies act to enhance the subduction rate of the STMW on the large-scale average.

This interpretation, however, is not applicable for small-scale eddies. The trajectories with instantaneous velocity field (Fig. 15c) look very similar to those with the unsmoothed monthly velocity field (Fig. 15b), indicating that the effect of small-scale eddies on the annual subduction rate is weak.

8. Summary

Using the results from a high-resolution general circulation model, this study provides a detailed description of the subduction process in the North Pacific. The annual subduction rate averaged over the whole North Pacific in the model is about 75% higher than that derived from the observed ocean climatology. This discrepancy may be attributed to the mixed layer parameterization and warm bias of the Kuroshio and its extension, a bias common to models that restore SST back to the observed climatology. We note, however, that the heavy spatial smoothing in preparing these climatological datasets weakens the MLD gradient and thereby reduces the lateral induction substantially.

Two marked peaks are found in the model subduction rate, in agreement with earlier observations of the mode waters (e.g., Talley 1999). One of these peaks corresponds to the STMW on the warmer side of the Kuroshio Extension, and the other to the CMW on the warmer side of the Oyashio Front. The maximum annual subduction rate associated with these two peaks exceeds 200 m yr−1. Sandwiched in between, roughly in the density range of 25.4–26.0 σθ, is a region where the annual subduction rate is rather weak, if not negative, with its magnitude generally <50 m yr−1. In this particular density range, the mixed layer has a strong seasonal cycle with its maximum deeper than 200 m in winter and minimum shallower than 25 m in summer in the western North Pacific, but the current velocity across the winter MLD front is rather weak, indicative of a weak lateral induction. This thus suggests that deep winter MLD is a necessary but not sufficient condition for the formation of mode waters.

The climatological data succeed in capturing the subduction peak for the STMW but fail to capture the peak associated with the denser CMW. The geostrophic flow computed from such a climatological dataset does not resolve the separation of the Kuroshio Extension and Oyashio Front (Fig. 16a), while a double-front structure is well defined in the model velocity field (Fig. 16b). This double-front structure is a commonly observed phenomenon (e.g., Joyce 1987; Schmitz et al. 1987), though most of the previous results were based on synoptic measurements. In a recent climatological analysis that averages the data along isopycnal instead of pressure surfaces on a 0.5° × 0.5° grid (Qu et al. 2001), this structure is also evident (Fig. 16c). We suggest that the double-front structure is critical for the subduction peak with the CMW.

For convenience we define eddies as perturbations from the zonally smoothed, annual mean field in this study. The large eddy contribution to the annual subduction rate is seen in the formation region of the STMW/CMW, amounting to up to 100 m yr−1. In a regional average (30°–44°N, 140°E–170W), this eddy contribution (12.4 Sv) accounts for 43% of the volume flux of waters entering the thermocline from the mixed layer (28.9 Sv). Most of this eddy contribution (9.7 Sv) is due to mesoscale eddies, while the effect of mean seasonal cycle (1.2 Sv) is negligible. Here we emphasize that the subduction of the STMW/CMW is a one-way process that allows mesoscale eddies to transfer more mixed layer water into the thermocline than bring it back from a Lagrangian point of view.

In this analysis, the gross effect of small-scale eddies (<30 days) is negligibly small, although they may locally modify the subduction process through their influence on the MLD and trajectories of water parcels. We note that the eddy-induced part of the mean flow—an eddy effect in nature (e.g., Hazeleger and Drijfhout 2000)—has been included in this analysis as the contribution by the mean flow. For this reason, the eddy-induced subduction rate may be underestimated in the present study. The contribution by the eddy-induced part of the mean flow can be estimated by comparing the present results with those from a coarse resolution version of the model, a subject we will leave for a future study.

Finally, we note that due to archiving limitations, the present analysis is based on five-day outputs of one model year alone. Because of this limited temporal sampling, we zonally smooth the model outputs to define the mean flow. It would be interesting to compare the results using temporal sampling if multiyear model output had been achieved. Also important to note is that the model used for this study was integrated only for 20 years because of its high computational cost. This might not be long enough for the northern subtropical gyre—especially structures of higher baroclinic modes—to reach a truly equilibrium state. As more powerful computers become available and allow a longer integration, our estimate of the annual subduction rate may be modified somewhat. But, in a qualitative sense, we believe that most of our conclusions will remain valid.

Acknowledgments

This research was partly supported by the Frontier Research System for Global Change through its sponsorship of the International Pacific Research Center and by NASA under JPL/NASA Contract 1216010. The authors are grateful to B. Qiu, J. McCreary, T. Yamagata, and K. Hanawa for many useful discussions. Thanks are also extended to L. Talley and two anonymous reviewers for thoughtful comments and useful editing on the earlier manuscript and to Y. Shen for constant assistance in using the model results.

REFERENCES

  • Cushman-Roisin, B., 1987: Subduction. Dynamics of the Oceanic Surface Mixed Layer, P. Muller and D. Henderson, Eds., Hawaii Institute of Geophysics Special, 181–196.

    • Search Google Scholar
    • Export Citation
  • Ebuchi, N., and K. Hanawa, 2000: Mesoscale eddies observed by TOLEX-ADCP and TOPEX/Poseidon altimeter in the Kuroshio recirculation region south of Japan. J. Oceanogr., 56 , 4357.

    • Search Google Scholar
    • Export Citation
  • Feng, M., H. Mitsudera, and Y. Yoshikawa, 2000: Structure and variability of the Kuroshio Current in Tokara Strait. J. Phys. Oceanogr., 30 , 22572276.

    • Search Google Scholar
    • Export Citation
  • Hanawa, K., 1987: Interannual variations in the wintertime outcrop area of the Subtropical Mode Water in the North Pacific. Atmos.–Ocean, 25 , 358374.

    • Search Google Scholar
    • Export Citation
  • Hanawa, K., and L. D. Talley, 2001: Mode waters. Ocean Circulation and Climate, G. Sieldler, J. Church, and J. Gould, Eds., Academic Press, 373–386.

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

  • Hazeleger, W., and S. S. Drijfhout, 2000: Eddy subduction in a model of the subtropical gyre. J. Phys. Oceanogr., 30 , 677695.

  • Hellerman, S., and M. Rosenstein, 1983: Normal monthly wind stress over the world ocean with error estimates. J. Phys. Oceanogr., 13 , 10931104.

    • Search Google Scholar
    • Export Citation
  • Hosoda, S., S-P. Xie, K. Takeuchi, and M. Nonaka, 2001: Eastern North Pacific subtropical mode water in a GCM: Formation mechanism and salinity effects. J. Geophys. Res., in press.

    • Search Google Scholar
    • Export Citation
  • Iselin, C. O. D., 1939: The influence of vertical and lateral turbulence on the characteristics of the waters at mid-depths. Trans. Amer. Geophys. Union, 20 , 414417.

    • Search Google Scholar
    • Export Citation
  • Ishida, A., Y. Kashino, H. Mitsudera, N. Yoshioka, and T. Kadokura, 1998: Preliminary results of a global high-resolution GCM experiment. J. Fac. Sci. Hokkaido Univ., Ser. VII (Geophysics), 11 , 441460.

    • Search Google Scholar
    • Export Citation
  • Joyce, T. M., 1987: Hydrographic sections across the Kuroshio Extension at 165°E and 175°W. Deep-Sea Res., 34 , 1333113352.

  • 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
  • Ladd, C., and L. Thompson, 2001: Water mass formation in an isopycnal model of the North Pacific. J. Phys. Oceanogr., 31 , 15171537.

  • Levitus, S., 1982: Climatological Atlas of the World Ocean. NOAA Prof. Paper No. 13, U. S. Govt. Printing Office, 173 pp and 17 microfiche.

    • Search Google Scholar
    • Export Citation
  • Marshall, D., 1997: Subduction of water masses in an eddying ocean. J. Mar Res., 55 , 201222.

  • Masuzawa, J., 1969: Subtropical mode water. Deep-Sea Res., 16 , 463472.

  • 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
  • Pacanowski, R., and S. G. H. Philander, 1981: Parameterization of vertical mixing in numerical models of tropical oceans. J. Phys. Oceanogr., 11 , 14431451.

    • 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.

    • Search Google Scholar
    • Export Citation
  • Qiu, B., and W. Miao, 2000: Kuroshio path variations south of Japan: Bimodality as a self-sustained internal oscillation. J. Phys. Oceanogr., 30 , 21242137.

    • Search Google Scholar
    • Export Citation
  • Qu, T., H. Mitsudera, and B. Qiu, 2001: A climatological view of the Kuroshio/Oyashio system east of Japan. J. Phys. Oceanogr., 31, 2575–2589.

    • Search Google Scholar
    • Export Citation
  • Schmitz, W. J., P. P. Niiler, and C. J. Koblinsky, 1987: Two-year moored instrument results along 152°E. J. Geophys. Res., 92 , 1082610834.

    • Search Google Scholar
    • Export Citation
  • Semtner, A. J., and R. M. Chervin, 1992: Ocean general circulation from a global edy-resolving model. J. Geophys. Res., 97 , 54935550.

    • 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. U.S.A., 76 , 30513055.

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

  • Talley, L. D., . 1999: Some aspects of ocean heat transport by the shallow, intermediate and deep overturning circulation. Mechanisms of Global Climate Change at Millenial Time Scales, Geophys. Monogr., No. 112, Amer. Geophys. Res., 1–22.

    • Search Google Scholar
    • Export Citation
  • 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
  • Williams, R. G., M. Spall, and J. C. Marshall, 1995: Does Stommel's mixed layer “demon” work? J. Phys. Oceanogr., 25 , 30893102.

  • Woods, J. D., 1985: The physics of pycnocline ventilation. Coupled Ocean–Atmosphere Models, J. C. J. Nihoul, Ed., Elsevier Science, 543–590.

    • 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 1958–1997: A GCM simulation. J. Phys. Oceanogr., 30 , 27982813.

    • Search Google Scholar
    • Export Citation
  • Yasuda, T., and K. Hanawa, 1997: Decadal changes in the mode waters in the midlatitude North Pacific. J. Phys. Oceanogr., 27 , 858870.

    • Search Google Scholar
    • Export Citation

Fig. 1.
Fig. 1.

Mean (a) and eddy (b) kinetic energy (cm2 s−2) of the model at 5 m

Citation: Journal of Physical Oceanography 32, 3; 10.1175/1520-0485(2002)032<0746:SOTNPM>2.0.CO;2

Fig. 2.
Fig. 2.

Depth (m) and density (σθ) of March mixed layer from (a) the model and (b) WOA98

Citation: Journal of Physical Oceanography 32, 3; 10.1175/1520-0485(2002)032<0746:SOTNPM>2.0.CO;2

Fig. 3.
Fig. 3.

Annual mean temperature (°C), salinity (psu), and potential density (σθ) vs depth (m) along 150°E from the model (left) and Qu et al. (2001: right)

Citation: Journal of Physical Oceanography 32, 3; 10.1175/1520-0485(2002)032<0746:SOTNPM>2.0.CO;2

Fig. 4.
Fig. 4.

Annual subduction rate and its two components by the mean flow: (a) vertical pumping, (b) lateral induction, and (c) subduction rate. Only positive values are plotted.

Citation: Journal of Physical Oceanography 32, 3; 10.1175/1520-0485(2002)032<0746:SOTNPM>2.0.CO;2

Fig. 5.
Fig. 5.

One-year Lagrangian trajectories of water parcels released at the base of March mixed layer calculated from the annual mean velocity field. Dots indicate grid points where the annual subduction rate by the mean flow is greater than 100 m yr−1

Citation: Journal of Physical Oceanography 32, 3; 10.1175/1520-0485(2002)032<0746:SOTNPM>2.0.CO;2

Fig. 6.
Fig. 6.

Same as Fig. 4 except for the total subduction rate consisting of contributions of both the mean flow and variabilities.

Citation: Journal of Physical Oceanography 32, 3; 10.1175/1520-0485(2002)032<0746:SOTNPM>2.0.CO;2

Fig. 7.
Fig. 7.

(a) Subduction rate (m yr−1) due to variability (referred to as the eddy-induced subduction rate in this study), and its two components due to (b) MLD variability and (c) velocity variability. Contour interval is 50 m yr−1, with an additional contour of 20 m yr−1 in (b)

Citation: Journal of Physical Oceanography 32, 3; 10.1175/1520-0485(2002)032<0746:SOTNPM>2.0.CO;2

Fig. 8.
Fig. 8.

Annual subduction rate induced by (a) mean seasonal cycle, (b) mesoscale eddies (>30 days), and (c) small-scale eddies (<30 days). Contour interval is 50 m yr−1

Citation: Journal of Physical Oceanography 32, 3; 10.1175/1520-0485(2002)032<0746:SOTNPM>2.0.CO;2

Fig. 9.
Fig. 9.

Annual subduction (Sv) per 0.2 σθ as a function of density. The shaded columns indicate the contribution of eddies

Citation: Journal of Physical Oceanography 32, 3; 10.1175/1520-0485(2002)032<0746:SOTNPM>2.0.CO;2

Fig. 10.
Fig. 10.

Potential vorticity (10−10 m−1 s−1) along isopycnal surfaces in March. The 200-m MLD contour is plotted in heavy solid line to mark the MLD front

Citation: Journal of Physical Oceanography 32, 3; 10.1175/1520-0485(2002)032<0746:SOTNPM>2.0.CO;2

Fig. 11.
Fig. 11.

Zonally smoothed (a) and unsmoothed (b) annual-mean meridional velocity (southward shaded; cm s−1) at 32°N from the model

Citation: Journal of Physical Oceanography 32, 3; 10.1175/1520-0485(2002)032<0746:SOTNPM>2.0.CO;2

Fig. 12.
Fig. 12.

Seasonal variation of the upper-layer (0–500 db) transport (Sv) through 32°N from (a) the model and (b) observations. In (b), the long-dashed line denotes the Ekman transport (Hellerman and Rosenstein 1983), the short-dashed line the geostrophic transport (relative to 2000 db; WOA98), and the solid line the total transport. Zonal integration starts from 143°E to 170°W to reduce the influence of the Kuroshio in the smoothed fields

Citation: Journal of Physical Oceanography 32, 3; 10.1175/1520-0485(2002)032<0746:SOTNPM>2.0.CO;2

Fig. 13.
Fig. 13.

Same as Fig. 11 except at 42°N

Citation: Journal of Physical Oceanography 32, 3; 10.1175/1520-0485(2002)032<0746:SOTNPM>2.0.CO;2

Fig. 14.
Fig. 14.

Same as Fig. 12 except at 42°N

Citation: Journal of Physical Oceanography 32, 3; 10.1175/1520-0485(2002)032<0746:SOTNPM>2.0.CO;2

Fig. 15.
Fig. 15.

Same as Fig. 5 except for (a) zonally smoothed monthly velocity, (b) unsmoothed monthly velocity, and (c) instantaneous velocity fields

Citation: Journal of Physical Oceanography 32, 3; 10.1175/1520-0485(2002)032<0746:SOTNPM>2.0.CO;2

Fig. 16.
Fig. 16.

Zonal velocity (cm s−1) and density (σθ) at 150°E calculated from (a) the WOA98, (b) the model, and (c) the climatological data prepared by Qu et al. (2001). Here geostrophic flow is referenced to 2000 db. Area with westward flow is shaded

Citation: Journal of Physical Oceanography 32, 3; 10.1175/1520-0485(2002)032<0746:SOTNPM>2.0.CO;2

 Table 1. Total subduction rate (Stot) and its components by the mean flow (Smean), mean seasonal cycle (Ssc), mesoscale eddies (Sme), and small-scale eddies (Sse) in Sv.

i1520-0485-32-3-746-t01

*

School of Ocean and Earth Science and Technology Contribution Number 5773 and International Pacific Research Center Contribution Number IPRC-98.

Save
  • Cushman-Roisin, B., 1987: Subduction. Dynamics of the Oceanic Surface Mixed Layer, P. Muller and D. Henderson, Eds., Hawaii Institute of Geophysics Special, 181–196.

    • Search Google Scholar
    • Export Citation
  • Ebuchi, N., and K. Hanawa, 2000: Mesoscale eddies observed by TOLEX-ADCP and TOPEX/Poseidon altimeter in the Kuroshio recirculation region south of Japan. J. Oceanogr., 56 , 4357.

    • Search Google Scholar
    • Export Citation
  • Feng, M., H. Mitsudera, and Y. Yoshikawa, 2000: Structure and variability of the Kuroshio Current in Tokara Strait. J. Phys. Oceanogr., 30 , 22572276.

    • Search Google Scholar
    • Export Citation
  • Hanawa, K., 1987: Interannual variations in the wintertime outcrop area of the Subtropical Mode Water in the North Pacific. Atmos.–Ocean, 25 , 358374.

    • Search Google Scholar
    • Export Citation
  • Hanawa, K., and L. D. Talley, 2001: Mode waters. Ocean Circulation and Climate, G. Sieldler, J. Church, and J. Gould, Eds., Academic Press, 373–386.

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

  • Hazeleger, W., and S. S. Drijfhout, 2000: Eddy subduction in a model of the subtropical gyre. J. Phys. Oceanogr., 30 , 677695.

  • Hellerman, S., and M. Rosenstein, 1983: Normal monthly wind stress over the world ocean with error estimates. J. Phys. Oceanogr., 13 , 10931104.

    • Search Google Scholar
    • Export Citation
  • Hosoda, S., S-P. Xie, K. Takeuchi, and M. Nonaka, 2001: Eastern North Pacific subtropical mode water in a GCM: Formation mechanism and salinity effects. J. Geophys. Res., in press.

    • Search Google Scholar
    • Export Citation
  • Iselin, C. O. D., 1939: The influence of vertical and lateral turbulence on the characteristics of the waters at mid-depths. Trans. Amer. Geophys. Union, 20 , 414417.

    • Search Google Scholar
    • Export Citation
  • Ishida, A., Y. Kashino, H. Mitsudera, N. Yoshioka, and T. Kadokura, 1998: Preliminary results of a global high-resolution GCM experiment. J. Fac. Sci. Hokkaido Univ., Ser. VII (Geophysics), 11 , 441460.

    • Search Google Scholar
    • Export Citation
  • Joyce, T. M., 1987: Hydrographic sections across the Kuroshio Extension at 165°E and 175°W. Deep-Sea Res., 34 , 1333113352.

  • 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
  • Ladd, C., and L. Thompson, 2001: Water mass formation in an isopycnal model of the North Pacific. J. Phys. Oceanogr., 31 , 15171537.

  • Levitus, S., 1982: Climatological Atlas of the World Ocean. NOAA Prof. Paper No. 13, U. S. Govt. Printing Office, 173 pp and 17 microfiche.

    • Search Google Scholar
    • Export Citation
  • Marshall, D., 1997: Subduction of water masses in an eddying ocean. J. Mar Res., 55 , 201222.

  • Masuzawa, J., 1969: Subtropical mode water. Deep-Sea Res., 16 , 463472.

  • 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
  • Pacanowski, R., and S. G. H. Philander, 1981: Parameterization of vertical mixing in numerical models of tropical oceans. J. Phys. Oceanogr., 11 , 14431451.

    • 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.

    • Search Google Scholar
    • Export Citation
  • Qiu, B., and W. Miao, 2000: Kuroshio path variations south of Japan: Bimodality as a self-sustained internal oscillation. J. Phys. Oceanogr., 30 , 21242137.

    • Search Google Scholar
    • Export Citation
  • Qu, T., H. Mitsudera, and B. Qiu, 2001: A climatological view of the Kuroshio/Oyashio system east of Japan. J. Phys. Oceanogr., 31, 2575–2589.

    • Search Google Scholar
    • Export Citation
  • Schmitz, W. J., P. P. Niiler, and C. J. Koblinsky, 1987: Two-year moored instrument results along 152°E. J. Geophys. Res., 92 , 1082610834.

    • Search Google Scholar
    • Export Citation
  • Semtner, A. J., and R. M. Chervin, 1992: Ocean general circulation from a global edy-resolving model. J. Geophys. Res., 97 , 54935550.

    • 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. U.S.A., 76 , 30513055.

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

  • Talley, L. D., . 1999: Some aspects of ocean heat transport by the shallow, intermediate and deep overturning circulation. Mechanisms of Global Climate Change at Millenial Time Scales, Geophys. Monogr., No. 112, Amer. Geophys. Res., 1–22.

    • Search Google Scholar
    • Export Citation
  • 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
  • Williams, R. G., M. Spall, and J. C. Marshall, 1995: Does Stommel's mixed layer “demon” work? J. Phys. Oceanogr., 25 , 30893102.

  • Woods, J. D., 1985: The physics of pycnocline ventilation. Coupled Ocean–Atmosphere Models, J. C. J. Nihoul, Ed., Elsevier Science, 543–590.

    • 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 1958–1997: A GCM simulation. J. Phys. Oceanogr., 30 , 27982813.

    • Search Google Scholar
    • Export Citation
  • Yasuda, T., and K. Hanawa, 1997: Decadal changes in the mode waters in the midlatitude North Pacific. J. Phys. Oceanogr., 27 , 858870.

    • Search Google Scholar
    • Export Citation
  • Fig. 1.

    Mean (a) and eddy (b) kinetic energy (cm2 s−2) of the model at 5 m

  • Fig. 2.

    Depth (m) and density (σθ) of March mixed layer from (a) the model and (b) WOA98

  • Fig. 3.

    Annual mean temperature (°C), salinity (psu), and potential density (σθ) vs depth (m) along 150°E from the model (left) and Qu et al. (2001: right)

  • Fig. 4.

    Annual subduction rate and its two components by the mean flow: (a) vertical pumping, (b) lateral induction, and (c) subduction rate. Only positive values are plotted.

  • Fig. 5.

    One-year Lagrangian trajectories of water parcels released at the base of March mixed layer calculated from the annual mean velocity field. Dots indicate grid points where the annual subduction rate by the mean flow is greater than 100 m yr−1

  • Fig. 6.

    Same as Fig. 4 except for the total subduction rate consisting of contributions of both the mean flow and variabilities.

  • Fig. 7.

    (a) Subduction rate (m yr−1) due to variability (referred to as the eddy-induced subduction rate in this study), and its two components due to (b) MLD variability and (c) velocity variability. Contour interval is 50 m yr−1, with an additional contour of 20 m yr−1 in (b)

  • Fig. 8.

    Annual subduction rate induced by (a) mean seasonal cycle, (b) mesoscale eddies (>30 days), and (c) small-scale eddies (<30 days). Contour interval is 50 m yr−1

  • Fig. 9.

    Annual subduction (Sv) per 0.2 σθ as a function of density. The shaded columns indicate the contribution of eddies

  • Fig. 10.

    Potential vorticity (10−10 m−1 s−1) along isopycnal surfaces in March. The 200-m MLD contour is plotted in heavy solid line to mark the MLD front

  • Fig. 11.

    Zonally smoothed (a) and unsmoothed (b) annual-mean meridional velocity (southward shaded; cm s−1) at 32°N from the model

  • Fig. 12.

    Seasonal variation of the upper-layer (0–500 db) transport (Sv) through 32°N from (a) the model and (b) observations. In (b), the long-dashed line denotes the Ekman transport (Hellerman and Rosenstein 1983), the short-dashed line the geostrophic transport (relative to 2000 db; WOA98), and the solid line the total transport. Zonal integration starts from 143°E to 170°W to reduce the influence of the Kuroshio in the smoothed fields

  • Fig. 13.

    Same as Fig. 11 except at 42°N

  • Fig. 14.

    Same as Fig. 12 except at 42°N

  • Fig. 15.

    Same as Fig. 5 except for (a) zonally smoothed monthly velocity, (b) unsmoothed monthly velocity, and (c) instantaneous velocity fields

  • Fig. 16.

    Zonal velocity (cm s−1) and density (σθ) at 150°E calculated from (a) the WOA98, (b) the model, and (c) the climatological data prepared by Qu et al. (2001). Here geostrophic flow is referenced to 2000 db. Area with westward flow is shaded

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 451 113 12
PDF Downloads 274 79 9