Near bottom currents and their relation to the transport in the Kuroshio Extension



[1] Near bottom current structure in the area of the Kuroshio Extension is examined using Lowered ADCP data obtained along 146°25′E and 152°30′E. The flow of the Kuroshio Extension reaches the ocean floor with a shift of the current to the right with depth. The near bottom currents have speeds of over 10 cm/s. Westward deep counter currents are observed to the north and south of the near bottom Kuroshio Extension. The eastward Kuroshio Extension transport is 163 Sv across 146°25′E and 113 Sv across 152°30′E, larger than previous studies because the contribution of the barotropic flow is twice that of the geostrophic flow. The downstream variation of the transport suggests the Kuroshio Extension strength increases after separating from the coast of Japan and decreases before 152°30′E. However, temporal variability could also contribute to the different transport estimates.

1. Introduction

[2] Extensive deployment of current meter arrays during the 1980s by Schmitz and collaborators [e.g., Schmitz, 1984; Schmitz et al., 1987] documented the eddy variability and the Eulerian mean flow in the Kuroshio Extension region of the North Pacific Ocean. The Eulerian mean picture is of an eastward flowing meandering jet in the upper ocean, with westward recirculation to the north and south of the Kuroshio Extension, and intense eddy variability in the upper kilometer. The abyssal flow was generally westward. At 165°E, the Kuroshio Extension and the recirculation to the north and south reach to the bottom [Joyce and Schmitz, 1988] but are only a few cm/s at 4000 m.

[3] Hall [1989] used a single current meter mooring to determine the velocity structure of the Kuroshio Extension at 152°E and found it reaches to the bottom with speeds greater than 5 cm/s and a southward shift of the current with depth. However, she did not examine the possibility of deep recirculation as known to exist in the Gulf Stream [Johns et al., 1995]. The estimated Kuroshio Extension transport was 86.9 ± 21.1 Sv but the observations did not sample the full width of the current and as a result missed perhaps 10 Sv from the warm side of the current.

[4] A synoptic section across the Kuroshio along the WOCE P10 line southeast of Japan (about 143°E) using CTD and lowered ADCP (LADCP) data demonstrated that the abyssal flow below the Kuroshio contributes to the volume transport [Wijffels et al., 1998; Qiu, 2002, Figure 5]. The total Kuroshio transport estimated was 143 Sv, larger than previous estimates [e.g., Hall, 1989; Teague et al., 1994; Niiler et al., 1985]. To the north and south of the Kuroshio there were strong, deep westward flows.

[5] The WOCE results demonstrated that, southeast of Japan, both the Kuroshio and the westward flows to its north and south are considerably larger than implied by the Eulerian mean picture obtained in the 1980s. Here we use recent detailed meridional hydrographic surveys (Section 2) across the Kuroshio Extension along 146°25′E and 152°30′E to examine this question downstream from the Wijffels et al. [1998] and Qiu [2002] results. We examine near bottom currents (Section 3) and demonstrate the importance of deep barotropic flows to the transport (Section 4).

2. Data and Methods

[6] Two spatially dense CTD/LADCP full-depth sections were occupied across the Kuroshio Extension (Figure 1). In May 2001, a meridional section was occupied along 146°25′E from 30°N to 36°45′N (ORV MIRAI, MR01-K04) when 28 casts were made with 28 km spacing. The LADCP used was an RD Instruments system (302 kHz) that was attached downward looking to the CTD frame. The velocity profile was obtained at a vertical resolution of 8 m. The LADCP could detect the sea bottom at a range of 200 m. In July 2000, another meridional section was occupied along 152°30′E from 30°N to 40°N (ORV MIRAI, MR00-K05). A total of 61 casts were completed with 18 km spacing. For most stations, data were obtained to within 10 m off the bottom. However at a few stations data were only obtained to within 70 m off the bottom and at one station within 200 m off the bottom as a result of restricted length of cable available. The same equipment was used on the two sections. During the occupation of the 152°30′E section, a current meter mooring was deployed at 152°30′E, 32°08′N.

Figure 1.

CTD/LADCP Observation lines in the Kuroshio Extension area superimposed on the ETOP5 topographic map and the climatological mean temperature field at 1000 m depth.

[7] The absolute velocity of the LADCP data over the ground was obtained by using the method of Cunningham et al. [2003]. That is, bottom-track data within 200 m above the sea bottom are used to calculate a vertical profile in 10 m bins from 20 m to 170 m from the bottom at each site (Figure 2). The near bottom currents, obtained by averaging the bins vertically, have a standard error of 2.0 cm/s. The LADCP data obtained along 146°25′E are of slightly lower quality than those along 152°30′E because only data from the upward casts were available along 146°25′E. There are three sites where the error exceeds twice the standard error. When these doubtful sites are not used, the standard error reduces to 1.8 cm/s on average, and is almost the same along 146°25′E and 152°30′E. The standard errors are independent for each station and lead to uncertainties of 17 Sv and 14 Sv for the barotropic transport based on the near bottom currents across each section.

Figure 2.

Latitudinal distribution of near bottom zonal velocity (top), and cross-section of zonal velocity and potential temperature (bottom) across (a) 146°25′E and (b) 152°30′E. In the top panel, the near bottom zonal velocity at each site is marked by a circle with the error bar indicating its standard deviation. The solid line indicates the smoothed velocity. In the bottom panel, the eastward (westward) velocity is red (blue). Speeds less than 5 cm/s are not colored. The temperature counter interval is 1°C. The grey shaded area indicates the depths over which the near bottom velocities were averaged.

[8] To focus on the non-tidal currents, we remove the tidal currents for eight primary (M2, S2, N2, K2, K1, O1, P1, Q1) and two long period (Mf, Mm) constituents estimated with the TPXO.6 tidal model [Egbert and Erofeeva, 2002]. The estimated tidal currents have amplitudes less than 2.1 cm/s during the observation periods and had little impact on the observed near bottom currents. The zonal component of the near bottom flow (Figure 2) shows small-scale spatial variability with a magnitude of 5 cm/s even after removal of the tidal currents. This variability is particularly evident between 30°N and 33°N along the 152°30′E section. Currents at 4000 m from the current meter mooring at 152°30′E, 32°08′N show similar variability that is probably associated with either unmodelled tidal motions or inertial oscillations. To minimize the impact of this variability, a boxcar filter is used to average the flow over 83 km along 146°25′E (3 stations) and 92 km along 152°30′E (5 stations). Bad data were not used and for the few stations where there is missing data. This amounts to an interpolation using surrounding stations. Geostrophic transports were then calculated with no smoothing of the CTD data and using the near bottom currents as reference values.

[9] The surface flow field is obtained by summing the Eulerian mean current [Uchida and Imawaki, 2003] and geostrophic velocity anomaly derived from gridded altimeter data of the ‘Maps of Sea Level Anomaly’ [Archiving, Validation, and Interpretation of Satellite Oceanographic Data, 1997].

3. Surface and Near Bottom Flow Structure

[10] The smoothed near bottom currents are shown in Figure 3, along with the surface flow field.

Figure 3.

Near bottom currents based on LADCP (thick red arrows), and surface flow based on satellite altimeter data (thin blue arrows) around (a) 146°25′E and (b) 152°30′E.

3.1. Flow at 146°25′E

[11] At 146°25′E, the surface Kuroshio Extension flows east across the section at 33°N to 34°N, then immediately turns northward and westward re-crossing the section at about 35°N before turning east again at 36°N. Along 146°25′E, the station spacing was somewhat larger than at 152°30′E. Strong near bottom currents with a magnitude of up to about ±15 cm/s are apparent, slightly stronger than the near bottom currents at 152°30′E and much larger than the estimated errors.

[12] The strongest eastward near bottom currents are in a broad region between 32°N and 34°N, below and to the south of the surface flow (Figure 2a). The westward surface flow at 35°N is connected to a deep westward flow at 35°30′N and the eastward surface flow at the northern limit of the section (36°45′N) is connected to a deep flow to the east at 36°20′N. That is, as the Kuroshio Extension crosses the section flowing to the east, west and then east again, there is an associated abyssal flow offset to the right (looking downstream) of the surface flow. In addition, immediately to the north of the Kuroshio Extension and south of where the current crosses the section flowing westward there is a deep westward counter current. To the south of the Kuroshio Extension there is a westward counter current from the surface to abyssal depths. There is also weak eastward near bottom flow around 30°15′N.

3.2. Flow at 152°30′E

[13] At 152°30′E, the Kuroshio Extension is flowing almost directly east between 34°N and 35°N (Figures 2b and 3b). There is also a surface flow to the north between 35°N and almost 38°N. The near bottom currents along 152°30′E show a strong (10 cm/s) eastward jet between 32°N and 34°30′N; i.e., below and to the south of the surface flow. Immediately to the north of the Kuroshio Extension, there is a strong (over 10 cm/s) westward counter current (35°N to 35°40′N) and there is a weaker (6 cm/s) westward flow south of 31°N. The magnitude of these currents of about 10 cm/s is much larger than that of the estimated errors.

[14] Though the near bottom current is relatively weak to the north of the northern counter current, alternating direction zonal flow is apparent; eastward flow around 36°40′N, and westward flow at 37°N to 38°N. The position of these flows almost matches surface mesoscale flow patterns and suggests the barotropic eddy structure in the northern part of the Kuroshio Extension region can reach to the bottom. However, the surface eddy structures around 38°30′N and between 31°N and 33°N do not have a counterpart near the bottom.

4. Volume Transport

[15] Volume transports integrated from the southernmost site are calculated for the total transport, the barotropic transport based on near bottom currents (i.e., the water depth multiplied by the eastward component of the near bottom velocities) and baroclinic transport based on geostrophic flow relative to the bottom (Figure 4).

Figure 4.

Latitudinal distribution of integrated volume transport (thick line), barotropic component based on bottom velocity (thin line), and baroclinic component based on geostrophic flow relative to the bottom (dashed line) along (a) 146°25′E and (b) 152°30′E.

4.1. Volume Transport Across 146°25′E

[16] At 146°25′E the near bottom currents are stronger than at 152°30′E. Between 30°N and 32°N, the southern counter current transports about 50 Sv westward. Both the barotropic and baroclinic westward transports are about 35 Sv but these occur in slightly different latitude bands. The eastward flow associated with the Kuroshio Extension occurs between about 32°N and 34°N, and north of 35°45′N. The observations did not extend far enough north to cover the total eastward transport of the Kuroshio Extension or the northern counter current. Volume transport of the southern crossing of the Kuroshio Extension is 163 Sv, although it is not clear that the current fully crossed the CTD section before turning north and west. The barotropic transport contributes 113 Sv and the baroclinic flow contributes about 50 Sv. The ratio, 2.2:1, is similar to that at 152°30′E of 2:1 (see below). The baroclinic transport is offset to the north from the barotropic transport. The deep counter current just to the north of 34°N has a westward barotropic transport of about 25 Sv. The baroclinic flow associated with the Kuroshio Extension is 91 Sv between 32°40′N and the northern limit of the section. In the upper few hundred meters, there is some uncertainty in the transport estimates (perhaps several Sv) because of the non-linear terms associated with the meandering path of the Kuroshio Extension; 5 Sv at 146°25′E and 2 Sv at 152°30′E, where radius of curvature is 50 km.

4.2. Volume Transport Across 152°30′E

[17] The southern broad counter-current transports 34 Sv between 30°N and 32°30′N, mostly from the barotropic component. The eastward flow of the Kuroshio Extension between about 32°40′N and 34°50′N is 113 Sv. The barotropic flow contributes 75 Sv and further to the north the geostrophic flow contributes about 40 Sv. It is noteworthy that the northern counter current appears only after consideration of the barotropic flow and it transports about 60 Sv westward at 35°N and a further 20 Sv between 35°30′N and 40°N. Almost all of the transport in the northern counter current is in the barotropic component.

[18] Total volume transport between 30°N and 40°N along 152°30′E is small, only 5 Sv, as the eastward baroclinic contribution of 26 Sv is almost canceled by the barotropic westward contribution, −22 Sv. The broad existence of barotropic westward flow agrees with earlier studies [e.g., Niiler et al., 1985]. However, the total transport is less than the barotropic transport error estimates of 14 Sv.

5. Discussion

[19] The spatially dense CTD/LADCP measurements indicate near bottom currents of speed over 10 cm/s associated with and approximately parallel to the surface flow of the Kuroshio Extension. The width of the near bottom flows is over 100 km in latitude and the core is offset to the right (looking downstream) of the surface flow by 50 to 100 km. The present results are consistent with the analysis of Hall [1989] and Wijffels et al. [1998] and suggest this near bottom flow is a permanent feature of the meandering Kuroshio Extension. Westward deep counter currents are observed both to the south and north sides of the near bottom flow. The southern recirculation across 152°30′E is broad and the current speed relatively weak. However at 146°25′E, it is stronger than that at 152°30′E. The northern counter current is narrow (50 km along 146°25′E for the southern crossing Kuroshio Extension, 90 km along 152°30′E) and is mostly barotropic. Similarities of the present results to the synoptic section analysis reported by Wijffels et al. [1998] and Qiu [2002] suggests these intense counter currents are also permanent features of the Kuroshio and its Extension.

[20] The Kuroshio Extension transport estimates from these synoptic sections referenced to direct, synoptic current observations are larger than earlier geostrophic estimates referenced to a fixed level, the bottom or to two-year current meter averages. However, the values are similar to the synoptic sections using a combination of CTD and LADCP observations of Wijffels et al. [1998] and the estimate of Hall [1989]. The Wijffels et al. estimate is a transport of 143 Sv across about 143°E compared with the present results of 163 Sv (transport of the southern crossing of the current) across 146°25′E and 113 Sv across 152°30′E. The 152°30′E results are slightly larger than the estimate of Hall of about 97 ± 21 Sv (after correction for the missing transport) but within the error bounds. Together, these results suggest the Kuroshio transport increasing after separation from the Japan coast to a maximum and then decreasing further east. While it is possible that these different transports are merely a result of temporal variability, an analysis of altimeter data by Qiu [2000] suggested a maximum in the strength of the Kuroshio Extension and its southern recirculation occurring at about 150°E, roughly consistent with the above results. This spatial variation in transport is most likely forced by eddy-mean flow interactions [e.g., Qiu, 2002; Hurlburt et al., 1996].

[21] The westward counter current south of the Kuroshio Extension is estimated at about 50 Sv at 146°25′E and 33 Sv at 152°30′E, compared with the larger estimate of Wijffels et al. [1998] of over 130 Sv. However, the Wijffels et al. observations are for a broader latitude range than the present observations.

[22] Estimates of geostrophic transport are 81 Sv across 143°E [Teague et al., 1994], 91 Sv across 146°25′E (this study), 66 Sv, 50 Sv, and 54 Sv across 152°E [Niiler et al., 1985] and 59 Sv across 152°30′E (this study). These geostrophic estimates are comparable and again suggest the transport increases after separating from the Japan coast and then decreases.

[23] The comparative weakness of the Kuroshio Extension transport and the counter currents in climatological sections are most likely a result of inability to detect the barotropic component of the currents in geostrophic calculations and the smearing of the instantaneous currents as the Kuroshio Extension meanders back and forth across current meters. If the strong near bottom flows highlighted here are a permanent feature of the Kuroshio Extension, there are important implications for observational and modelling programs if we are to accurately quantify and simulate the total transports.


[24] The authors are grateful to Captain M. Akamine of the ORV MIRAI, his crew and researchers joining the cruises for their devoted cooperation. Altimeter data set was provided by the Collect Localisation Satellites, Space Oceanography Division as part of the Environment and Climate EU AGORA (ENV4-CT9560113) and DUACS (ENV44-T96-0357). This work was initiated as a part of the Midlatitude Ocean Project of JAMSTEC, and is made under the collaboration between JAMSTEC and CSIRO.