An Observational Study of the Kuroshio in the East China Sea: Local, Regional and Basin-Wide Perspectives on a Western Boundary Current

The Kuroshio is the western boundary current of the North Pacific midlatitude gyre. An observational study of the Kuroshio was conducted using data collected in the East China Sea (ECS) north of Okinawa from December 2002 through November 2004 with an array of inverted echo sounders and acoustic Doppler current profilers. Using these data, Kuroshio velocity structure and transport time series were obtained. Net absolute transport ranges between 4 and 29 Sv and has spectral peaks at periods of 60, 15 and 11 days. The mean net absolute transport is 18.5 ± 0.8 Sv. In conjunction with these in situ measurements, satellite altimeter data were used to extend the Kuroshio transport time series back to 1993. Comparison with net Ryukyu Current transport southeast of Okinawa shows that their mean sum (24 Sv), is less than the mean predicted Sverdrup transport. Additionally, Kuroshio and Ryukyu Current transports are positively correlated, with 60 day lag, due to the effect of mesoscale eddies impinging on the Kerama Gap. Finally, .annually-averaged Kuroshio and Ryukyu Current transports correlate positively with the Pacific Decadal Oscillation (PDQ) index. The correlations, r, which are highest at zero lag, are 0.76 for the Kuroshio and 0.49 for the Ryukyu Current. The combined transport variation correlated with PDQ index variation is about 4 Sv. PDQ index is strongly negatively correlated with NCEP wind stress curl over the central North Pacific at


Abstract
The Kuroshio is the western boundary current of the North Pacific midlatitude gyre. An observational study of the Kuroshio was conducted using data collected in the East China Sea (ECS) north of Okinawa from December 2002 through November 2004 with an array of inverted echo sounders and acoustic Doppler current profilers. Using these data, Kuroshio velocity structure and transport time series were obtained. Net absolute transport ranges between 4 and 29 Sv and has spectral peaks at periods of 60, 15 and 11 days. The mean net absolute transport is 18.5 ± 0.8 Sv. In conjunction with these in situ measurements, satellite altimeter data were used to extend the Kuroshio transport time series back to 1993. Comparison with net Ryukyu Current transport southeast of Okinawa shows that their mean sum (24 Sv), is less than the mean

Preface
This thesis is written in manuscript format. It discusses the Kuroshio from three different perspectives organized into separate chapters. Chapters 1 and 2 have been published as journal articles, except for Section 7 at the end of Chapter 1.
Observations of the world's oceans have identified a common feature: intense, narrow jets exist along the basins' western boundaries. Quantifying heat, mass, and momentum transport carried in these western boundary currents is critical for understanding local processes as well as the oceans' role in global budgets. As with many ocean phenomena, measuring the flow in western boundary currents is a technological challenge due, in part, to the spatial and temporal scales involved.
The mid-latitude North Pacific western boundary current is the Kuroshio.
This current transports warm water northward along the basin-edge past Taiwan and China, and northeastward past Japan before separating from the continental shelf and flowing into the North Pacific as a free jet. Recent measurements of this current as it passes through the East China Sea, where it is called the ECS-Kuroshio, provide the first long-duration, high-resolution dataset from which an absolute transport time series can be deduced. These measurements, made with inverted echo sounders and acoustic Doppler current profilers, were an observational success and the results of the data analysis are reported in this thesis.
Chapter 1 discusses the ECS-Kuroshio from a local perspective. This chapter describes the instruments used in the East China Sea experiment, discusses v techniques used to convert acoustic-travel-time measurements to transport, and quantifies the behavior of the current during the 23-month observation period.
Mean-and time varying-transport and velocity structure are reported as is the annual transport signal. A supplement at the end of Chapter 1 discusses periodic variations in the ECS-Kuroshio using spectral analysis, and complex empirical orthogonal function analysis.
Chapter 2 presents a regional perspective of the Kuroshio and other transports in the western North Pacific. Satellite altimetry data are used to extrapolate ECS-Kuroshio transport in time, resulting in a 14-year ECS-Kuroshio absolute transport time series. This is analyzed together with a similarly determined Ryukyu Current transport time series. The role of eddies in causing transport variations is investigated. Finally, the data are used with previously published results to deduce a regional mean-transport picture.
In the final chapter, Chapter 3, the mid-latitude North Pacific western boundary current system is examined from a basin-wide perspective. A significant correlation is found between the western boundary current transport and the Pacific Decadal Oscillation index. The role of wind stress curl over the North Pacific in causing this correlation is investigated. Additionally, correlations between PDQ index and other regional signals are discussed. Tables   Table 1. Locations,

Abstract
Kuroshio velocity structure and transport in the East China Sea (ECS) were investigated as part of a 23-month study using inverted echo sounders and acoustic Doppler current profilers (ADCPs) along the regularly sampled PN-line. Flow towards the northeast is concentrated near the continental shelf with the mean surface velocity maximum located 30 km offshore from the shelf break (taken as the 170 m isobath). There are two regions of southwestward flow: a deep countercurrent over the continental slope beneath the Kuroshio axis and a recirculation offshore which extends throughout the whole water column. There is a bimodal distribution to the depth of maximum velocity with occurrence peaks at the surface and 210 dbar. When the maximum velocity is located within the top 80 m of the water column, it ranges between 0.36 mis and 2.02 mis; when the maximum velocity is deeper than 80 m, it ranges ~etween 0.31 mis and 1.11 mis.
The 13-month mean net absolute transport of the Kuroshio in the ECS is 18.5 ± 0.8 Sv (standard deviation, cr = 4.0 Sv). The mean positive and negative portions of this net flow are 24.0 ± 0.9 Sv and -5.4 ± 0.3 Sv, respectively.

Introduction
The Kuroshio is a western boundary current, serving as the return flow for the wind-driven circulation of the North Pacific Subtropical Gyre. Its source is the 1 North Equatorial Current which bifurcates off the Philippines . The northward flowing branch of this bifurcation, the Kuroshio, sometimes loops into the South China Sea before passing east of Taiwan. There, part diverts to the east forming the Ryukyu Current , which flows along the eastern side of the Ryukyu Island chain, while the remainder enters the East China Sea (ECS) over the Ilan Ridge (sill depth ~775 m ) forming the ECS Kuroshio. The Ryukyu Island chain separates the ECS from the Philippine Basin ( Figure 1). There is one deep channel in this chain, the Kerama Gap south of Okinawa, with a sill depth of about 1000 m . In the ECS, the Kuroshio flows mainly just seaward of the shelf break before leaving through the Takara Strait, which is divided into two sections by a seamount (summit ~320 m depth); the northern section reaches ~460 m depth while the southern section reaches ~1400 m .
East of the Takara Strait, the Ryukyu Current and the Kuroshio rejoin and flow northeastward south of Japan until they leave the coast as a free jet known as the Kuroshio Extension. A schematic of the Kuroshio path inside and near the ECS is shown in Figure 1.
Previous studies of Kuroshio velocity structure in the ECS have reported northeastward Kuroshio surface currents reaching speeds of 3.5 knots . A subsurface velocity maximum at about 200 m depth has been detected in some velocity sections northwest of Okinawa (e.g., Ito et al., 1995; and references therein). However, a subsurface maximum is absent in the mean absolute velocity section of Oka and Kawabe 2 (2003) and it is unclear whether this is related to instrument spacing or to the transience of the subsurface maximum. The Kuroshio lies over the slope and its position has a standard deviation of about 10 km (e.g., ) due to Kuroshio meanders inside the ECS. The amplitudes of these meanders are much smaller than the meander amplitudes south of Japan and in the Kuroshio Extension. A southwestward flowing countercurrent beneath the Kuroshio has been reported (e.g., Ito et al., 1995;James et al., 1999). Moored current-meter data China Sea, submitted to Journal of Geophysical Research, 2008). The countercurrent transport and structure, however, are not well resolved, since these are a few discrete point measurements.
Southwestward flow has also been reported between the Kuroshio and the Ryukyu Islands forming a recirculation (e.g., James et al., 1999). This flow is typically strongest at the surface where southwestward surface currents can reach 2 knots .
Previous studies of Kuroshio transport in the ECS include calculations for the period between 1986 and 1988 made from 39 hydrographic sections referenced with surface current data measured by ADCP and Geomagnetic Electrokinetograph (GEK)  the ECS have relied on averaging snapshots taken over many years. Here we present a 23-month time series of Kuroshio net absolute transport in the ECS. We also report on the time-and-space varying velocity structure of the Kuroshio over the last 13 of these months (when measurements were more complete) and calculate the corresponding positive and negative transport time series. In addition to quantifying the Kuroshio volume transports, we investigate their time variabilities as well as those of Kuroshio position and width.

Data
The primary data sources for this investigation are 11 IESs which were deployed in the Okinawa Trough region of the ECS for nearly two years and 2 ADCPs deployed nearby on the outer shelf for 7-13 months ( Figure 2,  Figure 2). b Pressures are mean pressures recorded at instrument locations: Al 0.5 m, A2 5 m, and Cl-6/Pl-5 1 m above the seafloor. c u is the downstream (rotated 38° clockwise from 0° True) velocity component measured nearbottom: Al 6 m, A2 15 m, and Cl -6 5lm, above the seafloor.
Horizontal spacing of the instruments varied from about 20 km beneath the main Kuroshio axis at the northwestern end of the line to 40 km at the southeastern end. Each P-line instrument (PIES) was equipped with a Digiquartz pressure 6 sensor, and each C-line instrument (CPIES) was equipped not only with the pressure sensor but also an RCMl 1 Aanderaa current sensor moored 51 m above the bottom. These instruments measured round-trip, bottom-to-surface acoustictravel-time (t), bottom pressure and bottom temperature every hour. In addition, the upstream instruments (C-line) made hourly measurements of current velocity and temperature 51 m above the bottom. The data set is complete except for the current record and about 1/3 of the pressure record at site C5. The velocity cross section and transport results discussed here focus on the C-line, however the t measurements from the P-line are used in the optimal interpolation mapping procedure described in Section 3.1.2.

ADCP
The portion of the Kuroshio which flowed in waters shallower than 550 m, near or over the shelf, was outside the region sampled by the IESs. For part of the 2-year IES deployment, this portion of the flow was measured with two bottommounted ADCPs on the shoreward extension of the C-line. Details of the velocity structure measured by the ADCPs are reported elsewhere (Lim, 2008 Details about the data processing, including removal of occasional "jumps" 8 (probably due to bottom-fishing boats dragging the IESs ), detiding pressure records and correction of pressure-sensor drift, are described in detail in a data report (Andres et al., 2005).
Current data were corrected for the local magnetic declination (roughly 50W) and rotated 38° clockwise so that U is the offshore cross-stream (x) velocity component and V is the downstream (y) velocity component. This rotation is consistent with the orientation of the PN-line ( Figure 2). For sea water, sound speed and specific volume anomaly (8) both depend only on temperature, salinity and pressure. As a result, in many regions 'tref can be used as a proxy for a water column's 8 profile (He et al., 1998). This relationship of 8 to 'tref and pressure is called the Gravest Empirical Mode (GEM) , and the GEM empirical lookup table typically resembles first-mode baroclinic variations in the water-column density distribution. In strong baroclinic current regions the GEM usually represents the pycnocline sloping across the current system and accounts for most of the observed variability in 8 (Willeford, 2001;Sun and Watts, 2001;. Other sources of variability, such as internal waves or bottomintensified topographic waves, can exist without being represented by the GEM. In many applications, a single GEM field . is used as a o-lookup table for a given region, but in the ECS, we constructed two separate fields: a localized GEM for a small region around C6 and a main GEM for the rest of the IES sites.

Obtaining velocity fields and transports
Detailed bathymetry data with I-minute resolution

Off-shelf-calculating the transport
The lowpass filtered 'tref data from the CPIESs and PIESs were gridded at 10 km spacing in x and y using optimal interpolation (01) (Bretherton et al., 1976;Watts et al., 1989) with an empirically determined correlation length scale for 'tref of 55 km and the assumption of horizontally non divergent flow. For flow over sloping topography, this assumption results in isopycnals which are cut off rather than compressed by the topography. Using the GEM relationship, a time series of C-line 8 cross sections was generated from the mapped 'tref· These cross sections were used with the thermal-wind equation, to calculate time series of the Kuroshio's downstream baroclinic velocity, v, and transport both referenced to 700 dbar ( Figure 4, upper line). Here z is in the upward vertical direction, g is gravity, p density, and f the Coriolis parameter. . In order to capture the transport flowing between the Ryukyu Islands and the most offshore instrument (C6 at x = 182.6 km), 1refwas OI mapped slightly beyond the instrument array. This 12.4 km extrapolation is over a short distance relative to the 1ref correlation length scale of 55 km. Upper line is the baroclinic transport referenced to 700 dbar. Lower line is the barotropic transport. Tick marks denote the beginning of each month.
In order to calculate the barotropic reference velocity (i.e., the absolute velocity at 700 dbar) and its corresponding transport across the C-line, one needs to know the horizontal pressure gradient on an. appropriate geopotential (level) surface. Since the depth of each instrument was not known with sufficient accuracy (because pressure sensors cannot deliver l-to-10 ppm absolute accuracy), time-averaged current data from the CPIES current sensors were used to reference or "level" the pressure data.
Because of significant deep geostrophic shear in the ECS, the following method was used to "level" the C-line instruments. For each instrument site, i, the mean downstream velocity, V;, at any level, is related to the mean pressure, P;, at that level by geostrophy, where the derivative is taken along the geopotential (level) surface. The velocity at the 700 m level for each instrument location can be written as the velocity measured by the current sensor plus the velocity difference between the depth, di, of the current sensor and the 700 m surface. Thus in the mean, for the ycomponent, where t is time, T the duration of the measurements, Ui the y-component of velocity measured by the current sensor, and the velocity shear in the last term is determined from the travel-time measurements using the GEM and Equation (1). (2) and (3) and OI with an empirically determined correlation length scale for pressure of 35 km, the mean (relative) pressure on the chosen geopotential surface at each site, P;, was calculated from the measured current velocity, v;(t), and acoustic-travel-time, r/t). By comparison of this calculated mean pressure with the mean of the measured pressure record, P ;, a leveling constant, LC, was determined for each instrument and added to the respective pressure record:

So using Equations
With a second OI step, this time using the current sensor measurements and the "leveled" pressure measurements, Pi = Pi + LC, the barotropic velocity and transport were calculated ( Figure 4, lower line). The approximation made here is that the barotropic velocity, which is defined in this work as the velocity at 700 dbar, is equivalent to the velocity at 700 m. The difference in velocity between 700 m and 700 dbar is very small, because both Llz and av/8z are small.

Shelf and upper slope -interpreting the ADCP data
The ADCPs were deployed for only the last 13 months of the 23-month period during which CPIES data were collected. Even when the ADCPs were operational, current velocities were not measured in the topmost 30 m and 179 m at sites Al and A2, respectively. In an attempt to establish reasonable estimates of the Kuroshio transport over the upper slope and shelf (i.e., shallower than 550 m) for the entire 23-month period, several extrapolation methods were attempted.
Two of these are spatial extrapolations (filling in the tops of the velocity profiles) and one is temporal (extending into the first 1 O. months of IES deployment for which no contemporaneous ADCP data are available). In all cases, once the velocity profiles were determined, they were mapped onto a regular grid with 10 km horizontal spacing by linear interpolation. Zero velocity was assumed 15 km shoreward of the shelf break. The gridded velocities were then integrated to calculate the transport over the outer shelf and upper slope between x = -15 km

Spatial extrapolation
The spatial extrapolation methods were 1) vertical extension with a GEM and 2) a horizontal smoothing method. These spatial extrapolation methods are described in the Appendix. Resulting time series of transport over the upper slope and shelf are plotted in Figure 5. It is encouraging that the vertical-extension and horizontal-smoothing methods generally resulted in similar transport time series even though the procedures are entirely different. The rms difference between the transports calculated by these two methods is 1.0 Sv. Occasionally, however, the vertical extension method gave higher shelf transports (e.g., around the beginning of July 2004 in Figure 5). During these times vertical profiles of velocity (figure not shown) generated by the horizontal smoothing method appear unrealistic, with velocities changing abruptly with depth. Also, during these times, the vertical extension method suggests there is a velocity maximum near A2. The horizontal smoothing method simply cannot reproduce situations where there is such a maximum between Al and Cl. Consequently, we choose the vertical extension method as the more realistic method of spatial extrapolation.
Velocities measured at Al were typically small (usually of order 0.2 mis or less) and thus the transport calculations are insensitive to Al measurements. The vertical extension method was attempted both with and without the Al velocity measurements as inputs to the 01 calculations, and the resulting transports are nearly the same ( Figure 5, blue and red lines, rms difference 0.6 Sv). By using only the A2 measurements we obtain a transport time series for 13 months instead of just the 7 months of Al operation. Net transport on the upper slope and shelf (i.e., x < 35 km) determined from ADCPs, from vertical extension method using data from Al and A2 (blue), A2 only (red) and from horizontal smoothing method (green). Tick marks denote the beginning of each month. In upper left comer, mean values are listed in their respective colors.

Temporal extrapolation
The majority (77%) of the Kuroshio mean net transport was measured by the CPIESs with the remainder (that on the upper slope and shelf) measured by the ADCP(s). In order to obtain an estimate of this flow for the first 10 months of CPIES deployment, during which the ADCPs were not operational, the position and strength of the Kuroshio were established using off-shelf velocities from the CPIESs. Then an empirically determined analytical shape formula was used to extrapolate the net transport over the upper slope and shelf. The method is described in the Appendix. Figure 6 shows that during the 13 months when ADCP data were available, the transport values calculated by this extrapolation agree reasonably well with 20-day low-pass filtered net transports determined from the ADCP data (rms difference = 1.4 Sv). This extrapolation of net transport enables the analysis of seasonal variability by extending the time series to nearly 2 years.
However, since the method oversimplifies the complex time-varying velocity structure of the Kuroshio, we use it to infer net transport only, and not to extrapolate the velocity structure or to calculate the positive and negative pieces of the net transport. Net transport on the upper slope and shelf(i.e., x < 35 km) determined from the analytic-shape method (thin line) and 20-day lowpass filtered transport from the vertical extension method applied to A2 ADCP data (thick line). Tick marks denote the beginning of each month. 13-month mean values are 3.0 Sv and 4.1 Sv for the analytic shape and 20-day lowpassed methods, respectively; the 23-month mean for the transport derived from the analytic-shape method is 3.1 Sv.

Results
In this section we report on the mean and variability of the observed velocity field and the transports.

Mean velocity structure
Using the CPIES and ADCP data as described in the previous section, C- There are two regions of mean negative flow: a deep countercurrent over the continental slope and a recirculation further offshore which extends throughout the water column. These features have been reported by other authors (e.g., Bingham and Talley, 1991;Ito et al., 1995;James et al., 1999;. Figure 7 and the near-bottom velocities measured by the ADCPs and CPIES current sensors (Table 1) show that, in the mean, the countercurrent at Cl and C2 does not extend over the shelf to Al and barely reaches A2 (mean at A2 is -0.3 cm/s 15 m above the seafloor and positive at higher elevations).
The mean velocity cross section along the PN-line reported by  (their Figure 3a) does not show vertical shear in the recirculation as strong as that in Figure 7. We verified that our vertical shear is not an artifact resulting from the use of two separate GEMs (the main GEM and the C6 "warm" GEM, Section 3 .1.1.) to generate o profiles. Even when a single GEM generated from all of the hydrocasts in the ECS (but including only the upper 700 m of the "warm casts") is used to determine o profiles, the shear in the recirculation is still present. In addition, calculations based only on the mean travel times at C5 and C6, without an 01 mapping step, result in strong vertical shear although in this case the shear is located between C5 and C6 rather than at the eastern edge of the instrument array. These tests indicate that while the vertical shear is real, its horizontal structure is uncertain since the instrument spacing on this end of the line is 40 km. Note, however, that the C-line cross section shown in Figure 7 has the P-line instruments' x-positions (see Table 1) superimposed on it (open circles).
Instruments from the P-line were used in 01 mapping 't (described in Section 3.1.2). Since this line, which is 40 km downstream of the C-line, falls within the empirically determined 't correlation length scale (55 km), the P-line data help constrain the location of this vertical shear.

Velocity variations
While the mean velocity cross section in Figure   In the mean cross section shown in Figure 7, the slope countercurrent is primarily confined between Cl and C2. The snapshots in Figure 8, however, show that the position of the countercurrent moves on-and offshore and occasionally reaches beyond C3, although the near-bottom velocity measured by the C3 current sensor is negative for only 6% of the 23-month record. Since the countercurrent width is comparable to the instrument spacing over the slope, our data do not adequately resolve this countercurrent.  suggested using the location of the l 8°C isotherm at 200 dbar as an indicator of ECS Kuroshio position (i.e., location of maximum surface 22 velocity). Our data show what previous authors (e.g.,  have pointed out, namely, that the maximum surface velocity often falls shoreward of the 200 dbar isobath, and thus the location of an isotherm at 200 dbar is not always a useful proxy for Kuroshio position. Guan (1980) defined the "core" of the Kuroshio as the region between the surface 0.4 mis isotachs. We have adopted this definition of the "core" and take its midpoint to be the "position" of the Zero contour is white. Contour interval is 0.2 mis. Black lines are edges of the Kuroshio "core" according to the definition of Guan ( 1980) and yellow line, mid-way between the black lines, is our "Kuroshio position." Distances are given relative to the shelf-break (170 m depth). Tick marks indicate beginnings of months.

Mean transport
From the full 23 months of measurements, the mean net absolute transport across the C-line is estimated to be 18.7 ± 0.5 Sv (cr = 4.0 Sv). This is very similar to the mean estimated from the 13 months for which the most reliable shelf transport values are available, 18.5 ± 0.8 Sv (cr = 4.0 Sv). Means are summarized in Table 2.
The mean net transport is about 3 Sv smaller than that reported by  for flow into the ECS east of Taiwan. Their 20-month mean absolute transport from September 1994 to May 1996 is 21.5 Sv. Also their maximum transport (33 Sv) and minimum (12 Sv) are both larger than the extrema observed in this study (Section 4.4). While this discrepancy in the means may reflect interannual variability in Kuroshio strength, it also may indicate that the system is not closed between the Ilan Ridge and the C-line with flow onshelf and/or leakage through the Ryukyu Island chain.  (1993), the 13 month mean "baroclinic" and "barotropic" (positive) transports across the entire Cline are 9.9 Sv and 14.1 Sv, respectively; across the 160 km measured by Cl and C6 they are 9.0 Sv and 10.9 Sv, respectively.
In a fine-resolution numerical model Guo et al. (2006)  fall onshelf transport is consistent with that inferred from observation . South of the PN-line, the model's onshelf transport is about 1.6 Sv (Guo et al., 2006). Observations indicate that flow is driven onshelf by a permanent cyclonic eddy north of Taiwan Wong et al., 2000) and by warm rings and filaments shed by Kuroshio meanders with wavelengths of 100-150 km and periods of 14-20 days propagating over the shelf-slope, observed in satellite infrared images  . Further evidence of upwelling onto the shelf related to the propagating meanders comes from towed ADCP measurements taken just southwest of the PN-line where horizontal confluence over the slope results in upward vertical velocities reaching 2.8 cm/s (Ito et al., 1995).
In Relative to their respective means, however, the negative transport variability is stronger by a factor of 2. Our mean positive transport is similar to that reported by , 23.7 ± 2.0 Sv (their 'KVT' which is the integral of all positive velocities along the section).
Net absolute transport measured by the CPIESs can also be split into baroclinic (referenced to 700 dbar) and barotropic components. The 23-month mean baroclinic transport measured by Cl through C6 (160 km of the C-line) is 12.8 Sv; the mean barotropic transport is 1.6 Sv. The barotropic component is only 11 % of the total mean transport. A fine resolution numerical model (Guo et al., 2003) also predicts a barotropic component (consistent with the definition used here) which is 11 % of the total transport. While this small barotropic component would seem to justify the assumption, made in previous ECS studies (e.g., Hinata, (1996)), that the baroclinic transport referenced to 700 dbar calculated from hydrography is a reasonable estimate of total transport, this is not necessarily the case at a particular time. As can be seen from the time series in Figure  the largest effect is in the recirculation whose magnitude is increased by 0.5 Sv.
There is a small component of transport in the ECS which is not captured here, namely that flowing across the 25 km between the southeastern end of the Cline and the Ryukyu Island chain. If the velocity goes to zero linearly from the end of the C-line to the island chain, we are missing about -1 .3 Sv of the recirculation transport. Some of this may be southward flowing water that has seeped through the island chain rather than recirculation of Kuroshio water in the ECS.

Transport variations
Time series of absolute transports across the C-line are shown in Figure 11.
The overall maximum net transport was 29.5 Sv on 28 April 2004, followed 38 days later by the overall minimum of 4.0 Sv on 5 June 2004. This may be related to a Kuroshio path transition to the large meander state south of Honshu, Japan, which occurred in July 2004 , Figure 7a).
Previous studies have reported that Kuroshio transport m the ECS is typically highest in summer and lowest in autumn (Fujiwara et al., 1987 andHinata, 1996, both reporting transports referenced to 700 dbar; Ichikawa and Beardsley, 1993, reporting absolute transport). These results were based on snapshots of the transport each season over many years. When the transports from the present study are averaged by season ( Figure 12, crosses), the highest transport is indeed in summer (19.6 Sv) but the fall transport (18.3 Sv) is the second lowest; the lowest transport (17.6 Sv) is in spring.  investigated the seasonality of the Kuroshio as it exits the ECS through the Tokara Strait by using sea-level difference from tide-gage data at Naze and Nishinoomote as a proxy for transport. For 19 years (1965)(1966)(1967)(1968)(1969)(1970)(1971)(1972)(1973)(1974)(1975)(1976)(1977)(1978)(1979)(1980)(1981)(1982)(1983) averaged by month, the maximum sea level difference (and by inference the maximum transport) was in July and the minimum in October. However, the character of seasonal signal varied between years and was split into four types by Kawabe: small-amplitude, semiannual, and annual signals with two different phases. Monthly averaged transports for the present study are shown in Figure 12.
When monthly mean net absolute transports are calculated as an average of all data values falling within a given month over the two-year period, the highest transport is in August (21.4 Sv) and the lowest is in October (15.9 Sv), which is close to the 19-year mean result reported by Kawabe. The seasonal signal in the data from the current study most closely (though not exactly) resembles Kawabe's semiannual type which has maxima occurring January to April and July to September and minima falling in April to June and October to November. Red dots represent monthly means for the first year; green dots represent monthly means for the second year; blue dots represent monthly means for the two years averaged together (these are plotted and joined over a two year period to show clearly the yearly cycle). += seasonal (3-month) means for the two years averaged together. Complete data for November and December are only available for the first year. Red and blue bands show periods ofKawabe's (1988) semiannual-type minima and maxima, respectively (see text).

Conclusions
Near the PN-line north of Okinawa, we have obtained transport and velocity cross section time series lasting nearly two years. These show net absolute transport varying between about 4 Sv and 30 Sv with a 19 Sv mean value.
While the maximum and mean transports are comparable to those reported previously, the minimum transport from this study is lower than any transport from previous studies  and references therein). The seasonal signal of transport from this study most closely resembles  semiannual-type with monthly mean transport minima in both years occurring in  Modeling by  suggests that the shape of the velocity cross section (e.g., position of surface velocity maximum and rate of velocity decrease with depth), rather than the magnitude of the transport, determines whether most of the Kuroshio leaves the ECS through the northern or southern section of the Takara Strait. This in tum is thought to influence the large meander state of the Kuroshio south of Japan .  study was performed using a shape formula from Xue and Mellor (1993) having a single velocity maximum at the surface. It would be interesting to determine the effect of a subsurface maximum on the modeled path of the Kuroshio as it exits the ECS.
This study confirms the presence of a persistent countercurrent beneath the Kuroshio. Notably the near-bottom velocity at site C2 measured by the CPIES current sensor is in the downstream direction for only 26% of the 23-month record and its mean value is -3 cm/s More densely spaced instruments are required to study the structure of this countercurrent since it is typically only 20-30 km wide.
While there is certainly a countercurrent present, in order to capture (1) nongeostrophic dynamics and (2)

Supplement -Periodic variability
Previous researchers have reported that Kuroshio position and transport fluctuate with variability concentrated in specific frequency bands. James et al.
(1999) used inverted echo sounders to study Kuroshio meanders in the ECS.
During their 14.5-month deployment, they found evidence of persistent meanders with 11-day period and intermittent meanders with 7-and 16-day periods.
According to their instability model, the 11-day wave corresponds to the period of the fastest growing instability.

33
Variability of the inflow to the ECS east of Taiwan during a 20-month deployment of moored ADCPs is reported by . The most energetic spectral peak is at 3-4 month period (~100 days) with secondary peaks at 35-and 15-day period. The 100-day peak is thought to be related to the arrival of mesoscale eddies from the ocean interior .
While causes of the 15-day variability are not discussed,  noted that the transport below the main thermocline showed this peak more strongly than the transport above this level.
Periodic variability of the Kuroshio in the East China Sea during our instrument deployment is examined with (1)  period. CEOF is analysis is performed using the the entire 23-month t records.
Results are compared with previous studies in the region and with ECS wind data.

Spectra
As described in Section 5, transport crossing the C-line in the ECS is  , and 11 days (11.1-11.7 days). Figure 15 shows the variance preserving spectrum of position (defined as the center between the 0.4 mis isotachs as in Section 5.2). Most of the energy is between 10-and 30day periods. Kuroshio position and net transport are only weakly positively correlated (r = 0.24 which is statistically significant at the 95% confidence limit).
Coherence calculations (not plotted), show that while position and net transport are not coherent at the 60-or 15-day periods, they are coherent near 11 days (squared coherence reaches 0.91 at 10.5-day period with transport leading position by 115.l 0 , which corresponds to 3.4 days) Net transport can be split into vanous components. Figure 16 shows found that their 15-day peak in inflow through the East Taiwan Channel was concentrated in the lower layer flow. The 11-day peak, is small in all spectra except the recirculation in which it is absent.
In order to isolate the variability occurring at these spectral peaks and determine if their amplitudes vary in time, transports are band-pass filtered ( Figure   18). In each case, the band-passing is done on that component of the flow in which a given peak is most pronounced: net transport is band-passed to isolate the 11day peak (panel a), lower layer transport is band-passed to isolate the 15-day peak   In all panels gray line is 2-day low-pass filtered transport and black line is band-passed signal. Panels show (a) net transport with 11-day band-pass (10-12), (b) lower layer transport with 15-day band-pass (13.5-1 6), and (c) positive transport with 60-day band-pass (45-70).

CEOF analysis
The 11-day and 60-day spectral peaks are further investigated usmg 't measured by the P-line instruments in conjunction with the C-line instruments to determine the horizontal structure of these waves. For each wave, we first bandpass filter the 11 -c records. Then we use CEOF analysis of the band-passed records to diagnose the waves, allowing for the possibility that they are propagating rather than standing waves. the variability is at short periods ( < 10-day period), there is long period variability in the area-integrated y-wind stress with a peak near 60-days, perhaps related to the Madden-Julian oscillation. Figure 23 shows the squared-coherence between the area-integrated wind stress curl and 't at instrument site Cl. The squaredcoherence at 60-day period is 0.66 which is significant at the 95% level. This coherence suggests that the winds may play a role in generating the 60-day wave locally, but the mechanism remains uncertain. Finally, we note the spectrum of transport crossing the C-line does not exhibit a ~ 100-day peak. It is not known why the ~ 100-day peak in transport detected by Johns et al. ( ) in 1994Johns et al. ( -1996

Spatial extrapolation by vertical extension
The first method for spatial extrapolation of the ADCP data uses a GEM lookup table (with 150 m streamfunction lf'1so, rather than 't700, as the lookup index) to obtain vertical extensions of the velocity profiles in the manner described below.
where<D~~~ = -J 8dp is the geopotential anomaly of the 150 dbar surface relative 700 to the 700 dbar surface. This is readily determined from 'tref using the GEM lookup table. The second term on the right-hand side of Equation 5 represents change in mass of the water column above the 700 dbar level. This is determined directly from the leveled pressure-sensor data.

I
The absolute velocity normal to the line of instruments is given by the gradient of the absolute streamfunction: Using V 150 at all sites and \f iso at the CPIES sites as input data, OI mapping is used to determine the time series of\f 150 at sites Al and A2 as described next.

Mapping the mean field and the residual
The OI mapping is done in two steps. Streamfunction and velocity are decomposed into the time-mean and a residual: where the overbar represents a time average over the duration, T, of the measurements.
The time-mean fields are mapped using a relatively large correlation length scale (85 km). The resulting mean fields are plotted in Figure 24. Note that the velocity plot is simply the first derivative of the streamfunction plot.
In the second OI step, the residual is mapped using a shorter correlation length scale of 35 km. This correlation length scale was determined from a plot (not shown) of correlation coefficient versus distance for pressure data with the common-mode signal removed. At each time, t, when \fi' 50 (x,t) has been 45 determined, the mean field, \f' 1so (x), is added to produce, by the first of these two equations, the time series of absolute streamfunction, \f' 1 so ( x, t) .

Lookup table
The next step in determining the velocity profiles at sites Al and A2 is to relate the streamfunction to the specific volume anomaly, 8. The absolute streamfunction can be separated into baroclinic and barotropic components: Assuming \f' 1 so ;:::: \f'bc gives This assumption is a potential source of error, but its validity was verified by good agreement between the shapes of those portions of the velocity profiles measured by the ADCPs (from 30 to 152 dbar and 179 to 285 dbar for sites Al and A2 respectively) with those calculated by the procedure described here.
From hydrographic data, a GEM lookup table relating 8 profiles and <!>;~~, is constructed. Using this lookup table, a time series of specific-volume-anomaly profiles is determined from the streamfunction time series.
Finally, the horizontal gradients in 8 profiles were used to determine vertical shears of velocity at sites Al and A2. These shears were then referenced with the V150 values, and the transport over the upper slope and shelf was calculated ( Figure 5, blue line) by integrating the velocities.

Spatial extrapolation by horizontal smoothing
The second method for spatial extrapolation of the ADCP profiles assumes that velocity varies smoothly along horizontal surfaces between Cl and Al, separated by about 41 km. The velocity profile . at C 1 was estimated from the CPIES, as described in Section 3.1.2. At Al a mixed layer in the upper 30 m of the water column was assumed and the shallowest velocity measurement was simply projected up to the surface. Velocity measurements at Al together with velocity and pressure at the CPIES sites were used as input to an OI mapping of the absolute stream function, 'l'i, at each level, i, in 10 m increments between 150 m depth and the surface. The OI procedure was done in two steps: first the mean fields and then the residuals (at 12-hour intervals) were mapped. Once 'l'i was mapped over the upper slope and shelf, the horizontal 'l'i gradient was used to 47 calculate an absolute velocity time series at A2 for that level. The A2 velocities calculated at each level between 150 m and the surface were then combined with the measured A2 velocities to obtain complete profiles. A2 velocities between 180 m (highest level measured by the ADCP) and 150 m (lowest level determined by horizontal smoothing) were linearly interpolated to avoid vertical discontinuity in the velocity profiles. The resulting transport time series is plotted in Figure 5 (green line). James et al. (1999) developed an analytical shape formula, based on that of Xue and Mellor (1993 ), to describe the Kuroshio basic state velocity cross section,

Temporal extrapolation using an analytical shape formula
Vc(x,z): x-X -z Z They empirically determined the constants (V0 , Ze, Xc0 , Zm, L) by fitting to a velocity cross section determined from hydrogtaphy referenced with surface ADCP data. Using this formula, an instantaneous Kuroshio velocity cross section, V(x,z,t), can be approximated by shifting this basic state shape horizontally a distance Xo(t) and multiplying it by a strength factor, S(t).
The position and strength of the Kuroshio were determined as follows. At each time, t, x0 (t), and S(t), were chosen such that [VLP (x,z,t) -S(t)·Vc (x-x 0 (t), z)]2 was minimized where VLP is the 20-day lowpass-filtered observed velocity (shorter period fluctuations were not well reproduced). This was carried out for the region between x = 35 km (location of CPIES 1) and x = 100 km near the edge 48 of the recirculation and between the surface and 400 dbar. Once the location and strength were determined for a given time in this manner, they were used with the shape formula and constants from James et al. (1999) to estimate velocities over the upper slope and shelf: Ve (x,z,t) = S(t)·Vc (x-xo(t), z), -15 m :S x :S 35 km.
The velocities were integrated to determine the net transport over the upper slope and shelf. Guo, X., Y. Miyazawa, and T. Yamagata (2006)

Introduction
According to Sverdrup theory, the North Pacific subtropical gyre transports water southward in response to the integrated wind stress curl. Convergence in this southward flow feeds the broad North Equatorial Current (NEC) which, in the mean, carries about 60 Sv westward across 137°E, roughly in accord with the transport expected from the basin-wide North Pacific wind stress curl . This current bifurcates east of the Philippines. The Kuroshio is the branch of the bifurcation that returns water northward as a swift western boundary current   . The Ryukyu Current is ephemeral near the southern-most Ryukyu Islands (e.g.,  and it is unknown how much of it comes from the Kuroshio east of Taiwan and how much from the ocean interior. There is evidence that the Ryukyu Current intensifies as it flows northeastward along the Ryukyu Islands (e.g., Nagano et al., 2008). South of Kyushu, the Ryukyu Current joins the ECS-Kuroshio as it leaves the Tokara Strait. Finally, the Kuroshio leaves the Japan coast as an eastward flowing free jet, the Kuroshio Extension, closing the North Pacific subtropical gyre.
The ocean regions east of Taiwan and the Ryukyu Islands are characterized by the frequent arrival of mesoscale eddies from the ocean interior at intervals of about 100 days Konda et al., 2005). These eddies originate in a zonal band of high eddy kinetic energy between l 9°N and 25°N and may be generated by baroclinic instability associated with the vertical shear between the shallow, eastward flowing Subtropical Countercurrent (STCC) and the underlying portion of the NEC . There is also evidence that some eddies may be generated by the passage of typhoons (e.g., .
Typical eddies, which can be cyclonic or anticyclonic, are about 500 km in diameter  with westward propagation speeds of 7-8 emfs (e.g., Konda et al., 2005), temperature anomalies of ±3°C, flow velocities around 20-40 cm s-1 and surface height anomalies around 20-30 cm (e.g., . Eddies hamper the ·evaluation of ECS-Kuroshio and Ryukyu Current mean transports from isolated hydrographic sections. Moreover, when these eddies arrive off Taiwan, they may change the proportions of transport flowing in the ECS-Kuroshio and the Ryukyu Current .
Previous studies of Kuroshio transport suggest that the annual range of variability on entering the ECS is less than 10 Sv , and references therein), which is small compared to that expected from non-topographic, time-dependent Sverdrup theory. The seasonally varying wind stress curl integrated over the Philippine Basin predicts a 20 Sv peak-to-peak annual range in Sverdrup transport (from  based on COADS data integrated from 125°E to 142°E) and that integrated over the entire North Pacific predicts about 50 Sv annual variation in transport their Fig. 1 based on climatological winds of ). An additional enigma is that North Pacific winds are strongest in winter (e.g., ) but the highest ECS-Kuroshio transport is typically observed in summer (e.g., . Sv annual variation of transport entering the ECS, they suggest that the Ryukyu Current should carry about 12 Sv in the mean, with a 16 Sv (peak-to-peak) annual range. Others suggest that, while Sverdrup flow does prevail in the interior , the reduced seasonal signal in the ECS arises from "JEBAR rectification" . To date, ECS transport studies have relied mainly on snapshots taken over many years (e.g., , or data collected quarterly . Thus, the month-by-month seasonal signal of the ECS-Kuroshio has not been well resolved. In addition, a long-term comparison between concurrent ECS-Kuroshio and Ryukyu Current transports has not been possible since the combined flow in these two Kuroshio branches has been reported only for isolated snapshots (e.g., . The role of eddies in steering the Kuroshio has not been thoroughly examined. The following questions are addressed in this paper.  In addition to the along-track SLA data, merged SLA maps at 7-day intervals from A vi so are used to track eddies in order to study their effects on the ECS-Kuroshio and the Ryukyu Current.

Ryukyu Current
We analyzed ECS-Kuroshio transport crossmg the C-line ( across the current, where ~SLA was determined with a combination of tide-gauge data and satellite-altimeter data. This technique provided a temporal extrapolation, resulting in a time series of NVT beginning in 1992 . Here we refer to this NVT as RT to distinguish it from the northeastward Kuroshio flow inside the ECS. Note that this RT had a strong eddy component. Its associated error was 5.9 Sv (2.1 Sv for the 10-month low-pass filtered values).

J2.4 Other regional signals
Absolute transport entering the ECS through the East Taiwan Channel (ETC) between Taiwan and the Ryukyu Islands was measured by Johns et al.

Methods
In this section, we first use C-line velocity cross-sections, determined from 2-day lowpass filtered in situ measurements taken hourly over 13 months (Andres et al., 2008), to demonstrate that sea surface height difference (~SSH) across the Kuroshio can be used as a proxy for full-water-column transport. Next we determine the empirical relationship between full-water-column transport from the in situ measurements and ~SLA from satellite altimetry. Finally, the resulting empirical relationship between these two is used with satellite altimeter data to generate a 12-year time series, beginning in 1992, of ECS-Kuroshio transport crossing the C-line.
In a layered ocean, assuming geostrophy, the net transport in the uppermost layer, VTsurf, is proportional to ~SSH across the current, gD

61
where g is gravity, D is the thickness of the uppermost layer, and f is the Coriolis parameter. Furthermore, if upper-layer transport is well correlated with full-watercolumn (total) transport, .1SSH can be used to infer the total transport. The seasonal signal due to surface warming and cooling, which is apparent in each of the two SLA time series, is assumed to be spatially uniform in this small region and thus cancels out in the ~SLA calculation. The net absolute ECS-Kuroshio transport across the C-line, determined from the in situ instruments (CPIES/PIESs and ADCPs), is also 40-day boxcar filtered (Figure 27d, heavy line). An empirical linear relationship between the resulting ~SLA and the transport is determined by least-squares fitting, and this relationship is used to calculate satellite-derived transport (Figure 27d, thin line). Net absolute transport, from the in situ instruments, and ~SLA (or satellite-derived transport) are well correlated with r = 63 II 0.83. Using this empirical relationship, rms difference between satellite-derived transport and transport determined from in situ instruments is 1.2 Sv. For the RyukyU Current crossing the 0-line, the correlation coefficient between satellitederived transport (from ~SLA) and that from in situ instruments was 0.91 with 2.8 Sv rms difference , although we note that the calibration procedure of  was slightly different from that used here (e.g., they smoothed the satellite data spatially, whereas we smoothed them temporally). To support the extrapolation method, the 12-year KT time senes is compared to .0.SSH across the Tokara Strait calculated using detided sea levels from Naze and Nishinoomote (locations are shown in Figure 25). Tokara Strait .0.SSH is a proxy for transport exiting the ECS through the Tokara Strait . In order to make the comparison, the Tokara Strait .0.SSH is KT is shown on the left and RT on the right. Dashed lines are 95% confidence intervals. Spectra are calculated with Hanning windows, 1100 days wide with 50% overlap (7 overlapping segments). Note the y-scales differ by a factor of 5.

Results and discussion
The 12-year mean KT is 18.7 Sv with± 0.2 Sv standard error, compared to 5.4 ± 0.4 Sv for the RT in the same period. Standard errors were calculated using the "blocking" method for corr.elated data described by . Transport of the ECS-Kuroshio is less variable than that of the Ryukyu Current with KT standard deviation about half that of RT (1.8 Sv vs. 3.9 Sv).
Variance-preserving power spectra of the two transports are shown in Figure 30.
KT shows two significant energy peaks, one at periods of 300-600 days and the other at 100-170 days. Also, a small, though significant, peak appears near 60 days. RT is broadly energetic at periods of about 100-500 days, with a couple of peaks between 100 and 200 days.

J 4.1 Annual cycle
Power spectra contain energy at the annual period in both the KT and the RT, although only the KT has a well-defined peak (Figure 30). In order to investigate the annual cycle further, transport data for the two currents are averaged by month over the 12 year time series for the two currents. These annual cycles and their sum are plotted in Figure 31 together with the standard deviation of transport for each month (dots). Even though KT is stronger than RT in the mean by a factor of 3.5, the annual range in RT is about 3.5 Sv, while that of KT is only about 1.6 Sv. Both currents show significant variability in the monthly transport, particularly from July to October (Figure 31 An October minimum m ECS-Kuroshio transport has been noted previously in other work (e.g., .
Moreover, 21 months of transport measurements just upstream of the ECS  averaged by month also exhibit an October minimum. Nevertheless, this minimum is absent in RT, which supports the possibility that the KT minimum results from local wind forcing  or the baroclinicity  rather than integrated wind stress over the Pacific or Philippine Basin.  found that transport in the ECS is positively correlated with the downstream component of wind stress.

Co-variation of ECS-Kuroshio and Ryukyu Current
KT and RT shown in Figure 28 are uncorrelated: r is not significantly different from zero. But this changes when one introduces a lag between the two currents ( Figure 32): with KT lagging RT by 60 days, the two currents become positively correlated at the 99% confidence level (r = 0.40). The coherence spectrum between KT and RT is plotted in Figure 33 (upper panel). Transports are coherent (at the 95% confidence level) at periods of 110 days, 160 days and 2-years. A 2-year spectral peak was also found by  who report a quasi-biennial oscillation in RT. In addition,  analysis of 18 years of sea-level difference across the Takara Strait has a spectral peak at 2.1 years. Phases at frequencies of significant coherence are plotted in Figure 33 (lower panel) and show KT generally lagging RT by 33-62 days. This suggests that the previously-noted strong correlation between the two ' currents occurring at 60 day lag .arises from processes occurring over a range of periods from about 100 days to 2 years. Further, the 60-day phase lag at 2-year period is relatively short, suggesting that the 2-year-period component in the ECS-Kuroshio and the Ryukyu Current is caused by the same driving force, possibly wind stress over the North Pacific .
There is an annual peak in coherence (Figure 33), but it falls below the 95% confidence level. It is significant only at the 90% confidence level (0.68).
Nevertheless it is worth noting that the phase of the annual peak differs by approximately 180° from that of the other coherence peaks, indicating that annual KT variation leads that of the RT by 137° ±13°, which corresponds to a lead of 4.6 months (or a lag of 7.4 months). This is consistent with the 5 month time difference in the monthly-average minima shown in Figure 31 (October for KT vs. March for RT).  Figure 33. Squared-coherence and phase lag spectra between KT and RT. For periods longer than 300 days, a 2220-day window was used. For shorter periods a 1100-day window was used. Gray lines on the coherence plot show the 95% confidence level. Phase (with standard error) is shown for frequencies which are coherent at the 95% confidence level. Additionally, the circled dot is the phase for 1-year period coherence (see text). Positive phase indicates RT lagging KT. Gray curves on the phase plot denote phase boundaries for transport lags between 30 and 70 days.

3 Eddy effects
The Kuroshio enters the ECS over the Ilan ridge between Taiwan and the Ryukyu Islands.   Current-meter data suggest there is a net mean flow through the Kerama Gap into the ECS . Generally, eddies themselves are not observed to pass through the gap into the ECS, but they appear to induce changes in the flow through the gap. The interaction of eddies with gaps has been treated numerically and analytically (e.g., McDonald, 2004, 2005) and with laboratory tank experiments (e.g., . Results depend on numerous parameters, such as the ratio of gap diameter to eddy diameter, Gld, relative position of eddy and gap, values off and fJ (=8f/8y), orientation of the boundary, whether or not there is background flow, and whether the problem is inviscid (numerical and analytical models) or includes friction (tank experiments). There is great "richness in dynamics," but, in tank experiments with Gld< 0.4, eddies are often observed to "funnel" water through gaps between cylinders without themselves passing between the cylinders     To discern the effect of eddies on ETC transport, KT, and RT, we tracked eddies using merged SLA maps available at 7-day intervals ( Figure 36). While ETC-transport shows many extrema during the period from October 1994 through January 1996 (e.g. Events i through vi in Figure 34b), only one of these (Event v, the minimum in October 1995) can be clearly related to an event affecting transport across the 0-and C-lines (Event II, track shown in Figure 36b ). Event I appears to decay near the Kerama Gap around June 1995 ( Figure 36a) and thus never reaches the region east of Taiwan. While Event III does eventually make it to the region east of Taiwan (Figure 36c), it does not arrive until May 1996, which is after Johns et al.'s (2001) ETC-transport measurements.
Five of the six pronounced extrema in ETC-transport appear to be related to eddies which never encounter the 0-line or the Kerama Gap (Figure 36d and e ).
The ETC-transport maximum in December 1994 (i), minimum in March 1995 (ii), and maximum in May 1995 (iii) appear to be related to eddies coming from the southeast (Figure 36d). ETC-transport maxima in September 1995 (iv) and December 1995 (vi) appear to be caused by anticyclonic eddies approaching from the east (Figure 36e ). For all six of these eddy-associated events at the ETC, there is a consistent pattern: transport maxima are associated with anticyclonic eddies and minima with cyclonic eddies, regardless of the direction of eddy arrival. This pattern is in agreement with the findings of  who based their conclusion on analysis of tide gauge-derived sea-level difference across the ETC, satellite altimetry and surface drifting buoys. There is one exception to this overall pattern: a weak ETC transport minimum in June 1995  appears to be due to the same anticyclonic eddy which eventually causes a transport maximum in September 1995 (Event iv). This unusual eddy has a large diameter and stalls for about 1 month in June 1995 near 124.5°N, 22.5°E (Figure 36e).
During that month the eddy appears to cause an offshore deflection of the Kuroshio away from the ETC resulting in the weak minimum in ETC transport.
This scenario is consistent with coincident drifter tracks  figure 9).
All three transport records shown in Figure 34 have strong oscillations with 3-5 month periods. In each case the oscillations appear to be caused by eddies arriving in the region from the east. Moreover a single eddy, like Event II, can be responsible for transport changes, at different times, in all three places, the 0-line, the C-line, and the ETC line. However, many eddies arrive at the ETC from the east or southeast rather than from the northeast. These affect ETC-transport without first influencing transports across the C-and 0-lines.  1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004  correlation coefficient and optimal lag are plotted in Figure 37. In each case, p (the probability that there is no correlation) is less than 0.05. From mid-1993 to mid= 1994 the optimal lag gradually drops from 100 to 50 days. From mid-1996 to mid-1997 the highest correlation coefficient is less than 0.4, implying that the Ryukyu Current transport crossing the 0-line and ECS-Kuroshio transport crossing the C-line are behaving somewhat independently. This may be related to anomalous ECS-Kuroshio transport during 1997-1998 when an El-Nino event occurred (Yuan et al., 2001). From 1998 to 2001 the optimal lag is 60 days; thereafter it drops to 40 days. We do not know what causes these changes in optimal lag, but they suggest that the propagation speed of eddies between the 0line and the Kerama Gap is variable. with transport passing through the Tokara Strait. Figure 38 shows C-line net transport (2-day low-pass filtered, determined with CPIESs and ADCPs). Also shown is net geostrophic transport relative to 700 dbar through the Tokara Strait which is determined four times per year from 12 CTD stations . The eight Tokara Strait transports measured during the CPIES/ADCP deployment are well correlated with the contemporaneous C-line transport (r = 0.87), and this correlation drops rapidly with lags (or leads) greater than one week implying that eddy-induced transport variations are advected rapidly inside the ECS. This is in contrast to the slow advection (~60 days) of eddies from the Oline to the Kerama Gap.

Comparison with regional mean flows and Sverdrup transport
The 12-year mean net transports crossing the C-and 0lines (KT and RT, respectively) sum to 24 Sv. This total transport is low compared to the throughflow across the ASUKA line south ofJapan (42 Sv, . But the location of leakage from the ECS-Kuroshio is not well constrained , and references therein)~ so it is not known what proportion occurs downstream of the C-line. Additionally, the preceding mass balance assumes that flow through the Ryukyu Islands between Okinawa and Amami-oshima is negligible; modeling by  suggests the exchange there is < 1 Sv.
Additional information about the mean flows in this region can be inferred using the results of   The inferred transports are shown as dashed gray lines in Figure 39.
This is consistent with a snapshot of the region based on 9 hydrographic sections throughout the region and an inverse technique . Their Figure 5 shows strengthening of the current between Okinawa and Amami-oshima which could be fed by interior flow, rather than recirculation. This intensification is qualitatively consistent with the 1/6° model of   and (2) Amami-oshima and the ASUKA line . The mean transports used to infer these flows are shown as black solid arrows: mean transport across the C-line (19 Sv, this study), 0-line (5 Sv, , ASUKA line (42 Sv, Imawaki et al.,200la), in the recirculation south of Japan (15 Sv, Imawaki,200la), across 24°N from the eastern boundary to 137°E (34 Sv, , and east of Amami-oshima (18 Sv, ). Other regional flows not used directly in these calculations are shown for reference as gray solid lines: Taiwan Strait (1 .2 Sv, , Korea/Tsushima Strait (2.6 Sv, , and between Taiwan and Yonagunijima (21.5 Sv, . Flow through the Ryukyu Islands south of Okinawa is not well constrained. Flow through the Ryukyu Islands north of Okinawa is assumed negligible here (< 1 Sv, .

Summary
Using satellite data calibrated with in situ transport measurements we have obtained a 12-year time series of net transport crossing the C-line (KT) and compared it to simultaneous transport crossing the 0-line (RT). Our study confirms what previous studies have suggested, that the annual variation in KT is smaller than that predicted by the non-topographic, time-dependent Sverdrup balance and the wind stress curl over the entire North Pacific. In addition, while the annual variation in RT is stronger than that of KT by a factor of 2, RT does not provide the 16 Sv annual variation predicted by  for transport calculated using the Sverdrup balance over just the Philippine Basin. Our study provides more observational evidence to support the model of  in which the annual range in transport is suppressed (relative to that predicted by Sverdrup theory) by the interaction of stratified flow with topography.
Despite the presence of the Ryukyu Island chain which shields the ECS from the Philippine Basin, arrival of eddies appears to cause much of the variability in the ECS-Kuroshio. The Kerama Gap, which is the deepest passage into the ECS, may be the conduit for communication of much of this eddyinfluence to the ECS even though the bulk of the Kuroshio transport enters the ECS near Taiwan. Maps of satellite SLA, together with positive correlation between RT and KT with ~60-day lag suggest the following portrayal. Eddies approach the Ryukyu Islands from the east with typical speeds of 7-8 cm/s (e.g., Konda et al., 2005). Their speeds are reduced to about 3-4 cm/s when they inducing flow through a gap) are consistent with numerical and laboratory studies . Some of these eddies then move to the southwest reaching the region east of Taiwan after many months, but their effect on the C-line transport is more clearly communicated through the Kerama Gap than through the ETC as demonstrated by comparison of correlation coefficients between the various transport time series. Once transport variations are felt along the C-line, they are advected quickly to the Takara Strait by the fastmoving ECS-Kuroshio. This conclusion is consistent with the findings of other authors who have noted a 60-day lag between the Ryukyu Current transport and sea surface height anomalies in the Takara Strait  and between sea surface height outside the Ryukyu Islands and volume transport through the Takara Strait . Hence the additional lag from the line where KT is determined to the Takara Strait must be small. Furthermore, modeling in the region  demonstrates the likelihood of communication through the Ryukyu Islands, particularly through the Kerama Gap, and  provide observational evidence with a salinity section through the Gap suggesting intrusion of water from east to west (their Figure 6b).
Finally, the combined KT and RT mean transports are less than the throughflow across the ASUKA line and less than the southward Sverdrup transport across 24°N from the eastern boundary to 137°E. These mismatches in mean flow each independently suggest that intensification of the Ryukyu Current downstream of Okinawa must be fed from the ocean interior rather than simply the recirculation south of Japan.

18. Introduction
Changes in Pacific climate occur on various time scales  and references therein). Interannual changes in sea surface temperature (SST) and sea-level pressure (SLP) are associated with El Nino/Southern Oscillation (ENSO; e.g., . Interdecadal Pacific climate variability is reflected in the Pacific Decadal Oscillation (PDO; . Here we investigate how the PDO manifests itself in the mid-latitude North Pacific western boundary current (WBC) system.

Previous researchers have developed proxies for Pacific interdecadal
variability from empirical orthogonal function (EOF) analyses of North Pacific monthly mean SST anomalies their NP index) and SLP anomalies . PDO index  is the time series of the first mode from EOF analysis of SST poleward of20°N from 1900 to 1993. When PDO index is positive ("warm phase"), SST is warmer than average near the west coast of America and cooler in the central North Pacific and the eastern North Pacific near the Kuroshio/Oyaspio region. Associated with this warm phase SST distribution are (1) lower than average SLP from 20°N to 60°N causing an enhanced Aleutian low , (2) positive wind stress curl anomalies centered on ~40°N, 155°W, and (3) negative wind stress curl anomalies centered on ~25°N, 170°W . During periods with negative PDO index ("cool phase"), this SST spatial pattern is reversed and the associated SLP distribution leads to wind stress curl anomalies which are negative at ~40°N and positive at ~25°N.

91
Ecological responses to interdecadal Pacific climate change have been well-documented. For example, Alaska salmon catch rates were low in the 1960s to the mid-1970s during a cool PDQ phase, and high from the mid 1970s to the mid 1990s during a warm phase their Figure 6). Also, biomass of large copepods south of Japan has been linked to decadal-scale climate variations, with high levels of biomass before the mid 1970s, low levels through the 1990s and high levels returning in the early 2000s  their Figure 3a).

Despite research documenting wide-ranging effects of PDQ on North
Pacific biology, the physical mechanisms underlying PDQ remain uncertain. It is not clear how observed SST changes in different ocean regions are interconnected.
Nor is it known how SST variability is related to changes in the ocean interior.  and

Study Area
Negative wind stress curl over the mid-latitude North Pacific sets up the subtropical wind-driven circulation: water flows southward, undergoes horizontal convergence giving rise to westward transport, and finally returns north as a WBC system. Figure 40 shows a schematic of currents in the western North Pacific.

Part of the WBC return flow enters the East China Sea (ECS) as the ECS-Kuroshio
where it is largely shielded from the ocean interior by the Ryukyu Islands. Inside the ECS, some of the ECS-Kuroshio feeds the Tsushima Current  which flows through the Korea/Tsushima Strait into the Japan/East Sea. The remaining ECS-Kuroshio exits the ECS through the Tokara Strait. Another part of the WBC system, the Ryukyu Current, flows outside the ECS along the eastern edge of the Ryukyu Islands, intensifying as it flows northeastward (e.g., . The Ryukyu Current and ECS-Kuroshio rejoin south of Kyushu and continue eastward. Some water recirculates in the Philippine Basin south of Japan as an anticyclonic gyre (e.g.,  and some passes over the Izu Ridge where it enters the North Pacific as a free jet, the Kuroshio Extension, south of which there is another recirculation cell.

Data
Transports at 10-day interval have been determined by calibration of satellite altimetry data with in situ measurements. Transports of the ECS-Kuroshio , Ryukyu Current , ASUKA-line throughflow (Imawaki et al., 2001a and b;, and Philippine Basin recirculation (Imawaki et al., 2001a and b) are available since the TOPEX-Poseidon (and later Jason-1) satellite began altimeter measurements in late 1992.
Daily sea-level data from tide gages at Nishinoomote, Nakanoshima, and Naze, corrected for the inverse barometer effect, are used to calculate both the sealevel difference (SLD) across the Tokara Strait (TSsw), which is a proxy for Kuroshio transport exiting the ECS , and the Kuroshio Position Index (KPI), which is a proxy for the location of the Kuroshio as it exits the strait .
Tsushima Current transport through the Korea/Tsushima Strait has been determined from SLD between Moji, Japan and Pusan, Korea, and bimonthly hydrographic data H. Na, personal communication, 2008).

95
The data were calibrated using transport determined from moored ADCP measurements taken during May 1999-March 2000 .
Wind data at ~2° horizontal resolution and 6 hour interval is available from the National Center for Environmental Prediction (NCEP; .
We use the data for 1993 through 2006 to calculate the wind stress curl field over the North Pacific.
The primary climate index used in this study is the PDO index . This is available through the Joint Institute for the Study of the Atmosphere and Ocean (JISAO) at the University of Washington.
We also make use of several other indexes. The Southern Oscillation Index (SOI), which is calculated from the SLP difference between Darwin and Tahiti, is provided by the NCEP Climate Prediction Center. Another ENSO indicator is the Cold Tongue Index (CTI), which is the average SST anomaly over 6°N -6°S and 180°W -90°W . This is available through JISAO.
The North Pacific Index (NPI), which is the area weighted SLP in the region 30°N -65°N, 160°E-140°W , is provided by the Climate Analysis Section, NCAR, Boulder, USA.

The connection between PDO and WBC transport
Pacific climate changed markedly around 197611977 ) causing a well-documented "regime shift" from a cool PDO phase to a warm phase (e.g., . The mid-l 970s change in PDQ phase may have been accompanied by a change in ECS-Kuroshio strength.   Although not explicitly stated by , the observation of low transport coincident with the pre-197 6 cool PDQ phase and high transport during the subsequent warm phase suggests a dynamical connection between PDQ and ECS-Kuroshio transport. This possibility is investigated here by comparing PDQ index with recently obtained ECS-Kuroshio and Ryllkyu Current transport estimates, paying particular attention to the 1998/99 change in PDQ phase. Figure 41c shows 14 years ofECS-Kuroshio transport determined from satellite altimetry calibrated with 13 months of in situ measurements . Similarly determined Ryukyu Current transport  is also plotted (Figure 41d). Superimposed on these 10-day interval transports are their 1-year running means. A positive correlation between the smoothed PDQ index and ECS-Kuroshio transport time series is clearly visible. Ryukyu Current transport also appears positively correlated with PDO index, although the agreement is not as strong, particularly around 1995.    1993 -2007, (c) 10-day interval ECS-Kuroshio transport, ( d) 10-day interval Ryukyu Current transport, ( e) daily SLD across the Tokara Strait, (f) daily KPI, and (g) transport through the Korea/Tsushima Strait. All transports in Sv (10 6 m 3 s-1 ).
One-year moving averages are superimposed on each plot. Tick marks denote the beginnings of the years. Years shown (inclusive) are (a) 1950-1992, (b-c) 1993-2007, (d-g) 1993-2004. The mid-1970s increase in ECS-Kuroshio transport documented by  is similar in magnitude to that observed here, and corresponded to the increase in PDO index in the "regime shift". Likewise, the 1993-2007 ECS-Kuroshio and Ryukyu Current transport changes presented here are in phase with the PDO index .. This is evident in lagged correlations of the transports with PDO index, which peak at zero lag ( Figure 43). In contrast, previous research (e.g.,   In regions with confidence levels lower than 90%, contour lines at 0.10 intervals denote regions of negative (positive) correlation in blue (red). Isobaths are shown at 500 and 3000 m depth.
The correlations noted above suggest the following scenario, consistent with Sverdrup theory. Enhanced westerly winds result in cooling of central North Pacific SST, either by increased atmospheric heat transfer or by southward displacement of the isotherms by Ekman drift (e.g., . This central Pacific cooling is a signature of the PDO "warm phase" (positive PDO index, with warming in the region off the coastal region of the northeast Pacific).
Enhanced negative wind stress curl also drives enhanced southward flow in the subtropical North Pacific gyre. This is compensated by a stronger return flow in the WBC system (higher ECS-Kuroshio and Ryukyu Current transports). Since correlations of WBC transport with PDO index and 'V Hx ' Z" w are strongest at zero lag, this appears to be a barotropic response to wind forcing.
This scenario leaves open the question of whether or not feedback from the ocean to the atmosphere plays a significant role in modulating the PDO. Based on results from a coupled ocean-atmosphere model,  suggested that the following feedback mechanism was important in setting the decadal timescale in Pacific SST.variability. Anomalously strong negative wind stress curl drives increased southward Sverdrup transport resulting in stronger WBC transport. This in tum transports more subtropical waters to the Kuroshio Extension region, decreasing the SST gradient there. The reduced SST gradient feeds back to the atmosphere and leads to reduced westerlies, which results in reduced Sverdrup transport and a reversal of the entire process.
In contrast to this model, in which heat transport by the WBC plays a central role,  concluded that the changes in Sverdrup flow and consequent changes in the WBC transport are not important in feedback associated with the PDO. This conclusion is based on Qiu's finding that changes in Kuroshio Extension strength are caused by westward propagating SSH anomalies forced by winds over the eastern Pacific (the speed of propagation is latitude dependent).
However, the WBC transport changes found in this study ( 4 Sv) are about half as strong as those Qiu found associated with westward propagating SSH anomalies (10 Sv), suggesting that advection of heat by the WBC may in fact be an important part of the PDO related atmosphere-ocean feedback loop.

Magnitude of the total wind-driven response
While the maps in Figure 44b-d confirm significant correlations of PDO index, ECS-Kuroshio and Ryukyu Current transports with V Hx Tw, they do not establish whether the magnitudes of the observed transport variations match the expected response to variations in the wind field. We address this question next by using the Sverdrup relation which is the steady state vorticity balance between meridional transport and wind stress curl, where Qy is the meridional volume transport anomaly per unit width (zonal), p is the density of seawater taken as 1030 kg/m 3 , f3 is the variation of Coriolis parameter,/, with latitude, df/dy, and '\! Hx T is the wind stress curl anomaly. For a given latitude, the total Sverdrup transport anomaly crossing a zonal section, Tsv, is Qy integrated over the length of that section. At each latitude, due to mass balance, WBC transport anomalies are expected to balance Qy integrated across the entire basin. Additionally, if the WBC response to the winds is rapid relative to changes in the wind field (e.g., if the ocean's response to the wind is transmitted via barotropic Rossby waves) Equation (13) can be used to determine a quasi-steady solution. Table 4 summarizes statistics comparing the observed transport anomalies with Tsv calculated as described in the following sections. These calculations indicate that there are components of the wind field, \7 Hx "fw, which (at zero time lag) are 1) not correlated with PDO index and 2) not reflected in the WBC transport observations.

Magnitude of the PDO-related wind-driven response
Next we isolate V' Hx Troo, that part of V' Hx Tw which is correlated with the PDO index. We then use the Sverdrup relation (13)   x  In order to determine which part of North Pacific V' Hx Tw is associated with PDO, annual mean wind stress curl, is regressed onto PDO index anomalies as follows. At each grid point in the North Pacific, V' Hx fw is averaged over year-long chunks, giving a 14-year record of annual means. Annual mean PDO index anomalies are calculated by subtracting the 14-year overall mean PDO index (0.296) from each annual mean PDO index. Then for each location, annual mean V' Hx Tw is regressed onto the annual mean PDO index anomalies, resulting in a map of regression coefficients. The resulting pattern in Figure 45 represents that part of the wind stress curl field associated with a PDO index anomaly of 1. Its spatial distribution is very similar to the 1st EOF of wind stress curl anomalies calculated by . Multiplying the pattern in Figure 46 by the PDO index anomaly time series gives a V Hx 'rrno time series; this extracts the part of the North Pacific wind stress curl field which varies in concert with PDO index. As PDO index increases (decreases), V H x 'rroo becomes more (less) strongly negative in a band between 20°N and 35°N and stretching from the western boundary to l 60°W. This is consistent with the positive relation noted previously between PDO index and WBC transport ( Figure   44), and with Sverdrup theory.
Using the PDO-related winds (V Hx 'rrno ) in Equation ( 13) Figure 47. Zonal cross-sections of North Pacific bathymetry. Panel (a) is a swath along 26°N . Shaded area shows region where the correlation between V' Hx rpoo and Ryukyu Current transport observations is significant (see Figure 44d). Panel (b) as in (a) for a swath along 24 °N with shaded area showing region of significant correlation between V' Hx rPDo and ECS-Kuroshio transport observations (see Figure 44c).

Comparisons with previous research
Kawabe (2001)  Kuroshio transport, and Ryukyu Current transport. They report a transport minimum in 1997 when we find relatively high PDO index and transports. Their high-transport years tend to coincide with high SOI (their Figure 8b) which is the opposite of our findings for the ECS-Kuroshio and Ryukyu Current (see Table 3).
PDO-like responses similar to ours may in fact be present off Taiwan, but being mainly barotropic, as suggested here and by the modeling of Kawabe (2001 )   although this is a complicated relationship (see for example Akitomo, 2007).
KPI, a proxy for Kuroshio position in the Takara Strait , is the ratio of the SLD anomaly across the southern part of the strait between Naze and Nakanoshima to the SLD anomaly across the whole strait between Naze and Nishinoomote ( Figure 49). The mean position in the strait is 29.9°N with standard deviation of 0.17° . Additionally, SLD across the Takara Strait between Naze and Nishioomote (TSsw) has been used as a proxy for Kuroshio transport .
Before exiting the ECS, 1-3 Sv ofECS-Kuroshio water intrudes onto the shelf, contributing to the northward flow transporting 1-4 Sv as the Tsushima Current through the shallow Korea/Tsushima Strait and into the Japan/East Sea . The transport of this intrusion varies seasonally and is highest in autumn . Based on geostrophic transport estimates and mass balance arguments, it is thought that the Tsushima Current and the ECS-Kuroshio annual mean transports may be anti-correlated  his Figure 29 panels band f). The seasonally varying ECS-Kuroshio is weakest in October  113 II which is when the Tsushima Current transport has been found to be highest . Contour lines at 200 m, 500 m, and 800 m depths. Stars show from north to south: Nishinoomote, Nakanoshima, and Naze. Arrows show alternative paths of the Kuroshio. The straight arrow is the "gap leaping" path, which results in the Kuroshio passing through the Tokara Strait in a southerly position (low KPI). The curved arrow shows the "topography-following" path in which the Kuroshio is constrained to flow through the Tokara Strait in a northerly position (high KPI).
A theory proposed by  to explain the observed anti-correlation between ECS-Kuroshio and Tsushima Current transports relies on the assumption of constant water volume above the subtropical thermocline, as proposed by . In this case, a relaxed therniocline tilt results in a weaker, but broader, WBC. This would result in less transport in the ECS-Kuroshio core but more flow on the shelf feeding the Tsushima Current.
This model is consistent with  observations of sea surface height determined from satellite altimetry (1992)(1993)(1994)(1995)(1996)(1997)(1998)(1999)(2000)(2001)(2002) and historical hydrocasts  in the Japan/East Sea. They found that sea surface height is negatively correlated with PDO such that during negative PDO phases, the surface water in the Japan/East Sea is warmer and saltier presumably due to the import of more subtropical waters.
The correlations listed in Table 3 show that the PDO index is positively correlated with ECS-Kuroshio transport and TSsw and negatively correlated with KPI and Korea/Tsushima Strait transport. These correlations suggest an alternative mechanism which does not rely on the assumption of constant water volume above the subtropical thermocline, but is still consistent with the correlation between sea surface height and PDO index noted by . The observed covariations may result from competition between topographic and inertial steering of the Kuroshio jet.  investigated the behavior of a WBC flowing past a boundary gap and found that potential vorticity conservation tends to make the current follow isobaths and hence loop into in the gap, while inertia tends to make the current leap the gap. As can be seen in Figure 49, in the ECS, along the shelf break, there is a kink in the isobaths at about 28°N. When inertially controlled, the jet continues in a straight path from this kink, it "jumps the gap" across the northern Okinawa Trough (Figure 49, straight arrow) and enters the Tokara Strait near its southern end (low KPI). On the other hand, when transport is relatively low, the jet is constrained to follow the topography, so it follows the shelf break and "loops" anticyclonically towards the Tokara Strait ( Figure 49, curved arrow), which it encounters at its northern end (high KPI).
When PDO index and ECS-Kuroshio transport both increase in response to wind stress curl (as described in the previous section) the ECS-Kuroshio may become 115 inertially controlled. This leads to ( 1) less transport intruding from the ECS-Kuroshio onto the shelf and lower Tsushima Current transport, (2) a more southerly position in the Takara Strait (i.e., lower KPI), and (3) increased transport exiting the ECS (higher TSsw).
Based on modeling and data,  and  correlate changes in KPI with changes in the current's background state and transport. They report that a highly stable southern path state in the Takara Strait (low KPI) is associated with an upstream cyclonic Kuroshio meander and thick inflow condition (large transport), while a weakly stable northern path state (high KPI) is associated with an upstream anticyclonic Kuroshio meander and a thin inflow condition (low transport). However,  suggests that the cause for the change in Kuroshio path state is a growing ECS-Kuroshio frontal meander such that a cyclonic eddy pushes the path from the northern to the southern state once it reaches the scale of the Okinawa Trough . This contrasts with our suggestion above that the jet position is controlled by the interplay of inertia and topography, however, the frequency range investigated by  and  is much higher (30-90 days) than the interannual variability investigated here.

PDO and downstream variations
After exiting the Takara Strait, the Kuroshio is rejoined by the Ryukyu Current, crosses the ASUKA-line south of Japan, and finally passes over the Izu Ridge. The ASUKA-line has been studied intensively and calibration of satellite 116 altimetry with in situ data (Imawaki et al., 2001 a, b) has resulted in transport time series of both the Kuroshio throughflow and the anticyclonic recirculation gyre ( Figure 40) in the Philippine Basin south of this throughflow . The westward flowing part of this recirculation is referred to as the Kuroshio Countercurrent by  and .
Since the ASUKA-line falls along a satellite altimeter track, the along-track The recirculation component is determined by assuming that its magnitude is j I! equivalent to the westward flow across the ASUKA-line between this point and a fixed point to the south, 26°N (Imawaki et al., 2001b). The difference between eastward and westward flow is taken as the throughflow. The mean throughflow is 42 Sv and mean recirculation is 15 Sv .
Despite the correlations between the ECS-Kuroshio and the Ryukyu Current with PDO index, we find that ASUKA-line throughflow is not significantly correlated with PDO index (Table 3). This is surprising, since the ECS-Kuroshio and Ryukyu Current presumably feed the ASUKA-line throughflow. ASUKA-line recirculation, on the other hand is negatively correlated with PDO index, though only at the 85% confidence level (r = 0.39), which suggests that as PDO index increases, recirculation strength decreases.
However, it may be that the procedure oflmawaki et al. (2001a and b) described above for splitting the eastward flow across the ASUKA-line into throughflow and recirculation components is flawed because of the use of a fixed point (26°N) to locate the offshore edge of the recirculation gyre. There is a shallow current, the Subtropical Countercurrent, which flows eastward south of the recirculation gyre . In the mean, this flow lies between 24°N and 27°N (Qiu and Joyce, 1992 their Figure 4). If the westward flow across the southern part of the ASUKA-line includes not only recirculation, but also interior flow feeding the western boundary current (i.e., the Ryukyu Current) and/or the Subtropical Countercurrent, then the ASUKA recirculation strength will be overestimated and ASUKA throughflow will be underestimated.
Additional evidence that the ASUKA-line calculations result in a misallocation of transport into recirculation and throuoghflow components comes from . He reports on a series of PIES and current-meter measurements along the ASUKA-line and along ~30°N from the ASUKA-line to the Izu Ridge in 2004 and 2005. The part of this array along 30°N was used to determine the recirculation strength: Using these data, Nagano then recalibrated satellite altimetry data, resulting in a time series of Kuroshio net transport which is quantitatively and qualitatively very different from that which relies on the fixed southern extent of the recirculation gyre.
Finally, there is some historical data surrounding the 1976/77 "regime shift" which support our expectation that increases in PDO index ought to be accompanied by increased Kuroshio throughflow south of Japan.  calculated the net transport south of Japan in this region in order to 118 investigate its relationship with the large meander south of Japan. Their net Kuroshio transport calculations (i.e., the throughflow), referenced to 1250 m, are based on twice yearly hydrographic surveys along 137°E from the Japan coast to 2°S taken by the Japan Meteorological Agency from the late 1960s through the late 1980s. They do not report the pre-and post-regime shift mean transports separately, but their Figure 7 shows that net transport from 1969-1974 is lower than that between 1977-1988. This is the sense of variation one would expect from increased PDO index correlated with increased ECS-Kuroshio and Ryukyu Current transport feeding increased throughflow south of Japan.

Conclusions
The use of in situ data to calibrate 15 years of satellite altimeter measurements has allowed us to establish that there is a positive correlation, at zero lag, between PDO index and WBC transport. This has been foreshadowed in the literature, but not established explicitly. Nitani (1.972) shows a figure (29d) with semi-annual transport relative to 1200 dbar, calculated by the Nagasaki Marine Observatory. From 1958 to 1961 this transport dropped, followed by a transport increase from 1962 through the late 1960s. This 1961 minimum is also apparent in the PDO index time series (Figure 4la).
PDO-related responses have also been reported in the Kuroshio Extension.
The magnitude of the WBC response to the PDO documented here is smaller than that of the Kuroshio Extension reported by  (about 4 Sv compared tol 1.6 Sv) and we find no evidence of a time lag between PDO index 119 and WBC transport. Thus the ECS response to PDO is different from that in the Kuroshio Extension. The data presented here suggest that interannual changes in II 1 the North Pacific WBC system arise from the integrated effect of the wind stress between the western boundary and the central North Pacific. Since correlations are highest at zero lag, the response is presumably barotropic.
The findings reported here do not preclude the suggestion by  that the PDO effect reaches the Kuroshio Extension region by baroclinic Rossby waves; both phenomena could occur. However, since the baroclinic effect of Qiu (2003) is much larger in magnitude (~10 Sv) than the barotropic effect discussed here (a few Sv), the latter could be difficult to detect in measurements taken in the Kuroshio Extension region.
Finally, we note that in the Philippine Basin there are alternating bands of eastward and westward flow (e.g., . Some westward flow I I I I 1 constitutes the anticyclonic recirculation south of the Kuroshio, while some is interior flow feeding the Ryukyu Current which, in the mean, triples in strength between Okinawa and Amami oshima (e.g. .
Due to the complicated flow pattern, in the Philippine Basin as well as the presence of many eddies which pass through the region ( e.g, , it is difficult to ascertain unambiguously from a single line of measurements which portion of the eastward flow is Kuroshio throughflow. The poor correlation between PDO index and ASUKA-line throughflow (particularly since the ECS-Kuroshio and Ryukyu Current are each well correlated with PDO