Journal of Oceanography, Vol. 54, pp. 681 to 694. 1998

681Copyright The Oceanographic Society of Japan.

Keywords:⋅K/O Zone,⋅ chemical alterna-tion/transport.

Chemical Alternation of Waters in the Kuroshio/OyashioInterfrontal Zone

T. ONO1*, I. YASUDA2, H. NARITA3 and S. TSUNOGAI3

1Graduate School of Fisheries Sciences, Hokkaido University, Hakodate 041-0821, Japan2Department of Earth and Planetary Physics, Graduate School of Science, University of Tokyo, Hongo 7-3-1, Bunkyo-ku, Tokyo 113-0033, Japan3Graduate School of Environmental Earth Sciences, Hokkaido University, Sapporo 060-0810, Japan

(Received 10 February 1998; in revised form 27 July 1998; accepted 8 August 1998)

An intensive survey has been conducted of the distributions of some chemical properties(dissolved oxygen, nutrients and carbonate properties) in the Kuroshio/OyashioInterfrontal Zone. Many low-salinity water patches were found down to depths of 640 m.Each chemical property also showed anomalies in these patches, but the degree ofvariation showed a low correlation with salinity. This may be due to the high variabilityof biological processes in the surface waters where these patches are formed. Verticalprofiles of the chemical properties were also observed along the Kuroshio extension axisfrom 140.50°E to 146.75°E. The concentrations of nutrients and total carbonate (TC) inthe water having densities greater than σθ = 26.60 can be regarded as being formed by theisopycnal mixing of the Kuroshio component water and Oyashio component water andbiological degradation within the density surfaces. This implies that the transport ofchemical properties by the diapycnal mixing is negligible in these density layers in theK/O zone.

1. IntroductionThe Kuroshio/Oyashio Interfrontal Zone (K/O zone:

Yasuda et al., 1996) is a key area when considering thetransport of heat, freshwater and chemical substances acrossthe subpolar front in the North Pacific. Complex mixing ofthe subtropical-gyre-originated Kuroshio water with thesubarctic-gyre-originated Oyashio water occurs in the K/Ozone and achieves effective cross-gyre transport (Talley,1993; Talley et al., 1995; Yasuda et al., 1996; Maksimenkoet al., 1997; Yasuda, 1997). In the subtropical gyre, NorthPacific Intermediate Water (NPIW) formed in the K/O zoneefficiently spreads the Oyashio-oriented substances fartherinto the gyre (Tsunogai et al., 1995). The transport of an-thropogenic carbon from the subarctic surface to the subtropicsubsurface carried by the newly formed NPIW is of particularimportance (e.g., Tsunogai et al., 1993; Tsunogai, 1997).

Processes relating to the transport of chemical sub-stances from the Oyashio to NPIW through the K/O zone arehighly complex. The density surfaces of NPIW (σθ = 26.7–27.0; Yasuda, 1997) and those of lighter densities lie in

shallow layers in the K/O zone, and the concentration ofeach chemical property in the newly formed NPIW may bedetermined not only by the isopycnal mixing but also bynon-hydrological processes within this area, such as bio-logical activities and gas exchange with the atmosphere.The contribution of each process in determining the chemicalcharacteristics of the new NPIW formed at the K/O zone hasnot yet been quantitatively evaluated. For the carbonateproperties such as total carbonate, even observed data arelimited for this region (Rogachev et al., 1996) despite itsobvious significance.

In spring 1996 we made intensive observations of thechemical properties including carbonate properties in the K/O region by several cruises. In this paper we describe thedistribution of chemical properties in the K/O region, lon-gitudinal changes in the chemical characteristics of theNPIW and other water masses along the Kuroshio extensionaxis and evaluate the contribution of each process affectingthese changes.

2. ObservationThe data were obtained during two cruises carried out

in May 1996, the WK9605 Cruise of the R/V Wakataka-Maru of Tohoku National Fisheries Research Institute(TNFRI) and the HK9605 Cruise of the R/V Hokko-Maru ofHokkaido National Fisheries Research Institute (HNFRI).

*Present address: Marine Productivity Division, NationalResearch Institute of Fisheries Sciences (NRIFS), Hukuura,Kanazawa-ku, Yokohama 236-8648, Japan.

682 T. Ono et al.

Table 1. List of observation lines used in this study.

Code Cruise Sampling location Frequency ofchemical sampling

Depth range Number of bottlesmeasured

A HK9605 (42°50N, 144°50E)–(39°00N, 146°45E) each 0.5°N 0–1000 m 13/stn.B WK9605 (34°51N, 140°50E)–(32°30N, 146°45E) 7 stns./line 0–1500 m 24/stn.147E WK9605 (39°00N, 146°45E)–(32°30N, 146°45E) each 0.5°N 0–1500 m 24/stn.149E WK9605 (35°01N, 149°31E) 1 stn./line 0–1500 m 24/stn.152E WK9605 (37°39N, 152°00E)–(36°20N, 152°00E) each 0.33°N 0–1500 m 24/stn.

Salinity (S), dissolved oxygen (DO), phosphate (P),nitrate (NO3), nitrite (NO2), silicate (Si), pH and total alka-linity (TA) were measured at each station. Salinity sampleswere measured by an Autosal located in the temperature-controlled room on board. DO subsamples were placed in100 ml Fram Bottles and measured with a HIRAMA ART-3DO-1 autotitrator within 6 hours of sampling. The averagedifference between measurement results of duplicate sampleswas 1 µmol/kg.

Nutrient samples were placed in 10 ml polypropylenetubes and stored at 4°C. Measurements were carried outwithin 3 days of sampling using a TRACCS-800 autoanalyser.Special grade reagents of Na2HPO4, KNO3, NaNO2, SiF6

(Kanto-Kagaku Co.) were used for primary standards of P,NO3, NO2 and Si, respectively. Concentrated standard so-lutions prepared in the laboratory were mixed and dilutedwith 3.5% NaCl solution on board and used as workingstandards.

Samples for pH and TA were placed in 120 ml poly-ethylene bottles and stored at room temperature. Measure-ment of pH was carried out within 12 hours of samplingusing a glass electrode calibrated with total hydrogen ionconcentration scale buffers (Dickson, 1993; DOE, 1994) ina 25°C water bath. TA measurements were carried outwithin 48 hours of pH measurements by the modified one-point method (Culberson et al., 1970). The average differ-ence between duplicate samples was 0.004 pHu for pH and3 µeq./kg for TA.

The total carbonate contents of water samples (TC)were calculated from pH and TA using the dissociationconstants of carbonic acid given by Dickson and Millero(1987) and that of boric acid given by Hansson (1973). On-board TC measurements were also carried out at 8 stationsby coulometry following the procedures of Ono et al. (1998).The ratio of measurement results of TC to the calculatedvalues at these 8 stations were 0.9963 ± 0.0036 on average.The precision of on board measurement of TC is below0.15% (Ono et al., 1998). Thus the variation in the ratio of0.36% (or approximately ±7 µmol/kg) is thought to revealthe overall uncertainty of the calculated TC values. In thefollowing discussions we use the calculated TC values only.

Observation stations were established along five observa-tion lines listed in Table 1 at approximately a 30 mileintervals. Locations of stations together with isotherm of100 m depth (TNFRI, 1996) during the observation periodare given in Fig. 1.

2.1 Measurements during the WK9605 CruiseWater samples were taken by CTD-RMS using 10L

Niskin Bottles. Samples from 24 depths were taken from theseasurface to 3000 m depth at each station. At severalstations where regional salinity minima were observed, thesampling depth was adapted so that water samples weretaken at each salinity minimum.

Fig. 1. Map of the observation stations. The 100 m watertemperatures during the observation period (TNFRI, 1996)are also shown.

Chemical Alternation in K/O Zone 683

2.2 Measurement during the HK9605 CruiseWater samples were taken by CTD-RMS with 1.7L

Niskin Bottles. Samples of 13 depths were taken fromseasurface to 1000 m depth at each station. S, P, NO3, NO2,Si, TA and total carbonate (TC) were measured at eachstation.

Salinity samples were placed in 500 ml glass bottles,stored at room temperature and measured by an Autosal onland. Nutrient samples were placed in 10 ml polypropylenetubes, stored frozen and measured on land using a TRACCS-800 autoanalyser.

Samples of TA and TC were placed in 120 ml glass vialbottles (Maruemu Co., Ltd.). Approximately 0.02 g of solidHgCl2 was added to each sample and stored at room tem-perature. TA samples were measured on land within 1month of sampling following the procedures as for WK9605.TC samples for HK9605 were measured by coulometry onland within 40 days of sampling following the procedure ofOno et al. (1998).

To evaluate the stability of the TC samples duringstorage, additional TC samples were collected at severalstations during the WK9605 Cruise besides those collectedfor on-board TC measurement. The results of on-and TCmeasurement 30 days after sampling agreed closely with theon-board measurements, with a difference of 3.1 ± 4.6 µmol/kg (on-board minus on-land) on average. Thus we concludethat the precision of TC measurements during the HK9605Cruise are about ±4.6 µmol/kg, and that the concentratechanges in TC samples during storage (3.1 µmol/kg on anaverage) was negligible within the measurement precision.

The preservation procedure of the TC samples describedabove has also been used elsewhere: Shitashima et al. (1996)preserved TC samples collected in the Philippine Sea for 4months using a similar preservation procedure. The watersamples collected during winter in the K/O zone have beenpreserved without significant change for over 100 days bythe same method (Ono, unpublished data).

3. Results and Discussion

3.1 Distribution of chemical properties along the 147°EsectionFigures 2a–2d shows the distribution of S, P, TC and

σθ, respectively, along the 147°E section (section codeslisted in Table 1). In the salinity section (Fig. 2a), a distinct,large low-salinity patch was found between latitudes 35.5°N–37.8°N with a core depth between 200–300 m. Under theKuroshio Extension (represented by the S = 34.80 isobar justsouth of 33.5°N) and just north of it, many small low salinitypatches were found, with major low-salinity peaks locatedat 34°N (150 m), 34°N (250 m), 33.66°N (400 m), 33.33°N(520 m) and 33°N (640 m). Some minor low salinity peaksalso occurred at 35°N (200 m) and 35°N (300 m).

These patch-like structures also occurred in both cross

sections of P and TC (Figs. 2b and 2c, respectively). Thedetailed location of each patch, however, seems to differsomewhat from the salinity profile. In the P section (Fig. 2b)the isobars show no clear patch between the latitudes 35.5°N–37.8°N and above 300 m depth. The signal also seems weakat 37.5°N in the TC section (Fig. 2c). Instead another patch-like structure is found at 39°N (500 m) in both P and TCcross sections.

Signals of patches at the circumference of the KuroshioExtension were also weak in Figs. 2b and 2c. No patches wasfound just under the Kuroshio Extension except for smallsignals located at 33°N (500 m) and 33.33°N (520 m). At thenorth side of the stream there seems to be some patch-likestructures similar to the structures shown in Fig. 2a. How-ever, the center of these patches shown in Figs. 2b and 2c arefound at around 35°N rather than 34°N, the center of thesalinity minima.

This disagreement in the distribution pattern betweenthe salinity and the chemical properties is also clear in thevertical profiles at each station. At 34.67°N, 146.75°E (Fig.3a), the AOU peak with a density of σθ = 26.80 seems verysmall, despite the corresponding salinity peak being thelargest of all the peaks observed at this station. The otherpeak, with a density of σθ = 25.79, seems remarkable in theAOU profile despite the fact that the salinity is rather faint.The AOU peak with a density of σθ = 27.08 is even positive,in contrast to other regional AOU peaks. At 33.53°N,146.75°E (Fig. 3b), a large low salinity peak is observedwith its center located at σθ = 26.57 in density. A corre-sponding peak is also observed in the AOU profile. Howeverits center is located at σθ = 26.42, which is significantlylighter in density than that of salinity.

3.2 Chemical characteristics of the low salinity patches inthe K/O zoneTo evaluate the chemical characteristics of the low

salinity patches quantitatively, the Oyashio content of eachpatch was calculated using both the salinity and phosphateconcentration as in the following equation;

Oy(x) = (Xm – Xk)/(Xo – Xk) (1)

where Oy(x) is the Oyashio content in each water patchcalculated from the property x, Xm the observed value of xat the core of the patch, Xk and Xo the values of x in each endmember of Kuroshio and Oyashio water at the densitysurface of the patch core, which ranged from σθ = 26.41 toσθ = 27.06, respectively. The x we show here is eithersalinity (s) or phosphate (p). (If we use nitrate, TC and TAinstead of phosphate, the resulting Oy(x) shows similardistributions as Oy(p).) Xk was calculated in each densityfrom the vertical profile observed at WK9605 Station 1(34.86°N, 140.83°E), the closest station to the Kuroshio axisoff Japan. Xo was calculated by averaging the profiles of the

684 T. Ono et al.

Fig. 2. Distributions of a) salinity (psu), b) phosphate (µmol/kg), c) total carbonate (µmol/kg) and d) potential density (σθ) along the147°E section.

Chemical Alternation in K/O Zone 685

Fig. 2. (continued).

686 T. Ono et al.

Fig. 3. Examples of vertical profiles of AOU (circles) and salinity (solid line) at a) WK9605 Stn. 17 (34.67°N, 146.75°E) and b) WK9605Stn. 19 (35.33°N, 146.75°E), respectively. The x-axis is the potential density (σθ), and the y-axis psu (left side) and in µmol/kg,respectively.

HK9605 stations in which the 100 m water temperature wasbelow 5°C. In calculating the Oy(p) we assumed that thewater mixing occurred mainly isopycnally in the K/O zone,and that the life times of the low salinity water patches in theK/O zone were short enough that the change in phosphateconcentration by biological processes in the water mass wasnegligible. We may be able to extend the above assumptions(and thus Eq. (1)) further for the wide range of watermasses

in the K/O zone, but in this section we would like to focusour interest on the chemical transport processes by the lowsalinity patches, the characteristic transport process ofOyashio component water through K/O zone. The overalltransport processes of chemical properties in the K/O zonewill be discussed in the following sections.

Figure 4 shows the correlation of Oy(p) and Oy(s) at 28patches with the density of 26.41 ≤ σθ ≤ 27.06 observed along

Chemical Alternation in K/O Zone 687

the 147°E section and B section. The standard deviation ofOy(p) around the regression line with Oy(s) is 0.47. Thismeans that if we assume a low salinity water patch with itsOy(s) of 0.5, this water mass can have any value of Oy(p)from 0 to 1. This implies that salinity gives no indicationabout the relative contribution of each patch in regard to thetransport of phosphate (or other chemical properties) fromsubpolar (Oyashio) to subtropical gyre water (Kuroshio)carried by it. The regression line itself, on the other hand, isOy(p) = 0.99 (±0.24)·Oy(s) + 0.38 (±0.12) (r2 = 0.41, n = 27).The bulk amount of the phosphate transport, integratedacross many water patches, thus seems to be proportional tothat of salinity, although the uncertainty in the regressionslope is rather large.

As to the cause of this low correlation between Oy(p)and Oy(s) as shown in Fig. 4, we can consider following twopossibilities: (1) Phosphate concentration changed after theformation of patches by biodegradation within the watermass. (2) There are some end members of the Oyashio water(with variable phosphate concentrations but the same salinity)in a density layer where the water patches are formed.

For possibility (1) we must conclude that this effect isnot the main source of the variation because almost half ofthe data in Fig. 4 are plotted above the Oy(p) = Oy(s) line.Even if the phosphate concentration changes significantlyby biological processes during the transport of water patches,such underwater biological activity (i.e., degradation oforganic matter) cannot produce water with Oy(p) smallerthan Oy(s).

For possibility (2), it seems possible that the phosphateconcentration varies significantly in hydrologically the same

water mass because of the strong biological activity in thespring K/O zone. The sea surface fCO2 distribution in thespring K/O zone, for example, clearly shows a large vari-ability with a finer horizontal scale than that of the salinityvariation (Nojiri et al., 1997). On the other hand Yasuda etal. (1997) showed that the low-salinity water patches in thisarea have two origins; i.e., patches having relatively lowpotential vorticity (PV) originating from the Okhotsk Seawater and those with relatively high PV originating fromopen Western Subarctic Gyre (WSAG) water. In the 147°Esection the patches observed at around 34°N, which areclearly observable in the salinity section (Fig. 2a) but lessclear in the P/TC sections (Figs. 2b and 2c), are found tohave a high density gradient (Fig. 2d) suggesting that theyare WSAG-originated high-PV patches. On the other handthose observed at around 35°N, which are clear in the P/TCsections than in the salinity section, seem to have a lowdensity gradient indicating that they are Okhotsk-originatedones. Similarly the patch observed at 37.5°N (Fig. 2a) seemsto have high-PV and that observed at 38.5°N (Figs. 2b and2c) seems to have low-PV. This implies that the Okhotsk-oriented low-PV patches are relatively more effective thanWSAG-originated ones when considering the transport ofbio-reactive (or atmosphere-oriented) chemical propertiesthrough the K/O zone carried by low-salinity water patches.

In this section we assume that the diapycnal mixing isnegligible compared to the isopycnal mixing in the K/Ozone. If the diapycnal mixing between a patch water and itscircumferences is significant, however, this may give anadditional source of variation of the data shown in Fig. 4. Toestimate the extent of this process more information is

Fig. 4. Scatter plot of Oy(p) against Oy(s) (for definition of the terms see text). Solid line indicates the regression line. The lineOy(p) = Oy(s) is also shown as a dashed line.

688 T. Ono et al.

needed about the hydrography of the upper layers in the K/O zone.

3.3 Change in the vertical profiles of chemical propertiesalong Kuroshio ExtensionBulk transport of chemical substances appears prima-

rily in the changes in vertical profiles of chemical properties

Fig. 5. Vertical profiles of a) density, b) salinity, c) phosphate and d) NTC in four Kuroshio extension axes from off the Boso Peninsula,Japan to 152°E.

at the axis of Kuroshio Extension. To investigate this theaveraged vertical profiles of density, S, P and TC at the axisof the Kuroshio Extension observed in each section (147°E,149°E, 152°E, and the Japan side axis of the B line, the axisin B line referred to as the “off Boso” axis) were calculated.Stations in which the water temperature was within 15–10°C at 200 m depth were selected for the calculation in each

Chemical Alternation in K/O Zone 689

Fig. 5. (continued).

section. In this calculation the TC values were normalized to35 psu with salinity (NTC). In the 149°E section chemicalsampling was carried out only at one station, so we use thisas the profile of Kuroshio Extension axis in this section. The200 m temperature at this station was 12.5°C and thus withinthe above criteria.

The results are shown in Figs. 5a–5d. The y-axis in eachprofile is shown as the potential density (σθ), with the ex-ception of the averaged density profile (Fig. 5a). The typicalprofile of each chemical characteristic in the Oyashio wateris also shown in Figs. 5b–5d. These Oyashio profiles arederived from the average values of the stations of the A line

690 T. Ono et al.

(Table 1) with 100 m water temperature less than 5°C. In theaxis east of 149°E each density surface with σθ smaller than26.60 seems to lie at depths shallower than the off Boso axis(Fig. 5a). This indicates the cooling of surface water withinthe Kuroshio Extension and/or K/O zone on the way of thewater flow.

The overall succession pattern of the salinity profilesfrom off Boso to 152°E is similar to that described in Yasudaet al. (1996). The profile with the salinity minimum of 34.2psu in the σθ = 26.90–27.00 density range (Type 1 water inYasuda et al., 1996) in the off Boso axis changes to a profilewith a minimum of 34.0 psu at around σθ = 26.70 density(Type 3 water in Yasuda et al., 1996) in the axis east of147°E. The salinity of the water with a σθ = 26.70–26.80density range in the axis east of 147°E approximates thevalue when the off-Boso axis water and typical Oyashiowater mix isopycnally with a 1:1 ratio. Note that each profilein Fig. 5b is of averaged data of several profiles for eachsection. Some stations have salinity profiles containingseveral regional salinity minima (Type 4 water in Yasuda etal., 1996) and consequently averaged profiles in Fig. 5boccasionally show small regional fluctuations.

In the density surfaces of 26.80 < σθ < 27.20 phosphateconcentration is higher in the Oyashio profile (Fig. 5c) thanin the Kuroshio profile because waters are older in thesubpolar region than in the subtropical region in thesedensity surfaces (e.g., Tsunogai et al., 1995). In Levitus andBoyer’s (1994) data set the phosphate concentration in theOyashio region is lower than that in the Kuroshio regioneven in the density surfaces of 26.80 < σθ < 27.20, but thismay be due to the difference in the sampling season in theirdata sources. The profiles east of 147°E lie roughly midwaybetween the Oyashio and the Kuroshio profiles, which is thesame feature as the salinity profiles. The phosphate con-centration on the density surfaces of σθ < 26.40, on the otherhand, tend to have smaller values than the off-Boso line inthe sections east of 147°E. (Fig. 5c). NTC profiles also showa similar pattern to that of phosphate (Fig. 5d).

The smallest concentrations of phosphate and/or TC inthe Oyashio profile exceed those of the waters with thedensity of σθ < 26.40 in the Boso line profile (Figs. 5c and5d). The decline of phosphate and TC concentration in eachσθ < 26.40 density surface east of the Off-Boso line thuscannot be explained by the mixing of Oyashio componentwater. Some mixed layer processes such as biological con-sumption of phosphate/TC are likely the course of thesenon-conservative concentration changes.

3.4 Non-conservative concentration changes in the chemi-cal properties of the Kuroshio extension water throughthe K/O regionYasuda et al. (1996) suggested that the water at around

σθ = 26.80 density in the Kuroshio Extension axis east ofJapan is formed by isopycnal mixing between the Oyashio

water and the off-Boso Kuroshio axis water. If we assumethat such isopycnal mixing processes occurs throughout thedensity surfaces, and that there are no non-conservativeprocesses in the K/O zone, the concentration of the chemicalproperty x in the Kuroshio extension axis east of 147°ECC(x) can be calculated by the following equation:

CC(x) = (1 – m)CK(x) +mCO(x)(m = (SK – Sm)/(SK – SO)) (2)

where CK(x), CO(x) and SK, SO the concentration of propertyx and salinity, respectively, in the corresponding densitysurface at the off Boso and the typical Oyashio axes, respec-tively. Sm the salinity of the corresponding water mass, andm the content of the newly added Oyashio water. CK(x) andSK were calculated in each density from the vertical profileobserved at WK9605 Station 1. CO(x) and SO were calcu-lated by averaging the profiles of the HK9605 stations inwhich the 100 m water temperature was lower than 5°C.

In the density surfaces of σθ < 26.40 we can not use Eq.(2) because typical Oyashio profiles occur only in the 26.40≤ σθ density range (Figs. 5b–5d). We thus assume that thereis no water mixing process between Kuroshio and Oyashiocomponents in these density surfaces; i.e.,

CC(x) = CK(x). (3)

Equation (3) well explains the salinity profiles of the Kuroshioextension waters in the density range of σθ ≤ 26.00 but notso well in the σθ = 26.20 density surface (Fig. 5b). We thushave not calculated CC(x) for this density.

Finally, the difference of actual concentration of x,Cm(x), from the CC(x) are then calculated; i.e.,

dCX = Cm(x) – CC(x). (4)

Stations in which the water temperature was within 15–10°C at 200 m depth were selected for the calculation ofCm(x) in each section. If our assumption about the watermixing is consistent, this reveals the amount of non-con-servative concentration change of properties at each densitysurface within the K/O zone.

Tables 2a and 2b shows the calculated dCX of variouschemical properties for both 149°E and 152°E axes, respec-tively. In the 149°E axis, dCP and dCTC becomes nearly zerowithin the observation error in the density surfaces of σθ ≥27.00 and σθ ≥ 26.90, respectively. In the 152°E axis againdCP and dCTC are nearly zero in these density ranges. Thissuggests that our assumption about water mixing is appro-priate and that these properties are nearly conservative in theK/O zone in these deep layers. This feature is also consistentwith the fact that the integrated transport of phosphate bylow salinity patches is proportional to that of salinity (Fig.4).

Chemical Alternation in K/O Zone 691

Table 2. Non-conservative concentration changes of chemical properties along the Kuroshio extension.

a) off Boso—149°E

Density Sal (psu) AOU P N Si TA NTA TC NTC NTC-6.6N

24.80 –0.051 –34 –0.15 –3.01 1.0 3 8 –25 –20 025.00 –0.057 –38 –0.16 –4.10 0.8 3 10 –26 –20 725.20 –0.013 –33 –0.14 –3.63 0.7 0 5 –20 –15 925.40 0.011 –35 –0.16 –3.68 –0.2 0 2 –22 –20 425.60 –0.013 –38 –0.16 –3.54 –1.0 4 7 –32 –30 –625.80 –0.021 –37 –0.15 –3.05 –1.2 7 9 –31 –29 –926.00 –0.025 –34 –0.17 –2.76 –4.7 –1 –1 –30 –31 –1226.2026.40 –0.21 –4.05 –7.9 –9 –3 –39 –39 –1226.60 0.03 –2.05 –1.3 4 2 –3 –10 326.70 –0.03 –3.54 –3.9 –4 7 –4 –16 826.80 –0.07 –1.95 –5.2 2 3 –7 8 2126.90 –0.09 1.43 4.9 1 3 3 4 –627.00 –0.02 0.19 –0.8 –2 1 1 3 127.10 0.01 1.11 0.6 –1 0 3 3 –527.20 0.04 1.11 2.2 2 4 7 7 027.30 0.01 0.48 –2.5 4 4 0 0 –327.40 0.04 0.77 –2.2 1 1 2 2 –3

b) off Boso—152°E

In the shallow layers, phosphate shows significantnegative dC values in the density surfaces of σθ ≤ 26.40 inboth 149°E and 152°E sections. Such significant negativedC values are also found in other properties in the shallowdensity surfaces (Tables 2a and 2b). The average values of

dC in the density ranges of σθ = 24.80–26.00 in the 149°Esection are –36 µmol/kg for AOU, –0.16 µmol/kg for P, –3.4µmol/kg for N and –26 µmol/kg for TC, respectively.

There are some potential uncertainty in the dC valuesdue to the uncertainties in the determination of end mem-

Density Sal (psu) AOU P NO3 + NO2 Si TA NTA TC NTC NTC-6.6N

24.8025.0025.2025.40 –0.004 –46 –0.11 –3.21 2.7 11 11 –29 –29 –825.60 0.049 –59 –0.18 –4.19 0.8 14 11 –37 –40 –1225.80 0.054 –63 –0.21 –4.55 –0.7 13 10 –38 –42 –1126.00 0.010 –79 –0.30 –5.43 –6.1 7 6 –43 –44 826.2026.40 –0.48 –8.27 –15.1 –11 –10 –61 –62 –726.60 0.08 –1.89 1.4 –4 –2 –16 –17 –426.70 0.02 –2.41 –1.8 0 1 –7 –8 826.80 –0.04 –1.74 –2.6 3 1 –5 –6 526.90 –0.08 0.73 2.9 1 –2 1 0 –427.00 0.00 0.18 –1.2 –1 –4 1 1 027.10 0.04 1.02 0.8 2 –2 7 6 027.20 0.09 1.10 2.1 1 –1 10 10 227.30 0.10 1.29 0.6 2 0 7 7 –127.40 0.06 –0.09 –2.8 1 2 5 5 6

>27.00 av. 1± 2.9

692 T. Ono et al.

bers, CK(x) and CO(x), and this must be considered in addi-tion to the simple observation errors. For CO(x) there is apossibility that the Okhotsk-oriented low PV waters (Yasuda,1997) is the significant source of Oyashio component waterin the K/O zone in addition to the WSAG-oriented typicalOyashio water. In our dataset low PV waters are observed atHK9605 Stn. 2, the northernmost station of the A line. If weuse the data of this station when calculating CO(x) instead oftypical Oyashio profiles, the resulting dCP decreases by 0.1µmol/kg in each density surface of 26.40 ≤ σθ < 26.90 in theboth sections. In the remaining density surfaces, dCp vary by±0.04 µmol/kg due to the change of CO(x). The variation ofdCTC due to the same operation is within the range of ±3µmol/kg for all sections and densities.

It is difficult to evaluate the potential uncertainty due tothe CK(x) calculation, because WK9605 Stn. 1 is the onlystation observed within the off Boso Kuroshio axis (Fig. 1).If we use WK9605 Stn. 2 (34.50°N, 141.49°E) instead Stn.1 for CK(x) calculation dCp and dCTC increase by 0.1 µmol/kg and 10 µmol/kg, respectively, in each density surface ofσθ < 25.90 in the both sections. In the remaining densitysurfaces dCp and dCTC vary by ±0.04 µmol/kg and 5 µmol/kg, respectively.

Finally, to evaluate the uncertainty due to the Cm(x)calculation in Eq. (4), one more station in the 152°E lineperipheral to the original ones are added to the Cm(x) datasource. The resulting dCp and dCTC varied by ±0.04 µmol/kg and ±5 µmol/kg, respectively, in each density surface.

To sum up the above arguments, the potential uncer-tainty of dCp due to the selection of end members are ±0.04µmol/kg in the density surfaces of 26.40 ≤ σθ and ±0.1 µmol/kg in the density surfaces of σθ < 26.40, respectively.Similarly the potential uncertainty of dCTC due to the se-lection of end members are ±5 µmol/kg in the density sur-faces of 25.90 ≤ σθ and ±10 µmol/kg in the density surfacesof σθ < 25.90, respectively. The uncertainty in the densitysurfaces of σθ < 25.90 might be overestimated, however,because the uncertainties in these densities originate mainlyfrom the uncertainty of CK(x) and the extent of this is notproperly estimated (see previous discussion). Isopycnaldifferences in the phoshate concentrations within theKuroshio axis should be smaller than the difference betweeninside the Kuroshio axis and outside it.

The decline of phosphate and TC in the density surfacesof σθ < 26.00 in the 149°E line (Table 2a) is thus concludedto have some significance. The potential uncertainty ofother properties were also evaluated and significance ofconcentration changes in these light density surfaces deter-mined. If a body of water flows through the Kuroshioextension directly, the residence time of this body in the K/O zone will be only a few weeks. In the case that the waterbody is incorporated in warm core rings, however, this bodyremains in the K/O region over a few years until the ringdeclines (e.g., Yasuda et al., 1992). The observed large dC

values in the shallow layers will thus be mainly sustained bythe mixed layer processes occurring within such long-livedwater bodies.

Silicate in each axis also shows significant negative dCvalues in the σθ = 26.00–26.40 density layers. However inshallower density surfaces dCSi becomes almost zero, re-gardless of the properties of the other nutrients. This can beexplained by the fact that the density surfaces of σθ ≤ 26.00outcrops mainly in the area south of the subpolar front in theK/O region (e.g., Talley, 1988), where diatom bloom is notso strong as that in the subpolar gyre.

In the density surfaces with the density of 26.40 ≤ σθ <26.90, it is difficult to conclude whether each dC value iszero or not when considering the potential uncertaintiesdiscussed previously. The overall trend shows that the dCsof each property have extensive minus values in the densitysurfaces of around σθ = 26.40, and approach zero with in-creasing density. If these negative dC values are real, thismeans either that there are some non-conservative concen-tration changes (i.e., biological degradation) in these den-sity surfaces or that the diapycnal mixing occurs amongthese density surfaces.

If the negative dC values are mainly caused by thebiological degradation, the corresponding dCN and dCTC

shows a Redfield relationship; i.e., if we calculate followingequation, dCred must be around zero.

dCred = dCTC – 6.6·dCN. (5)

The coefficient of dCN in Eq. 5 was derived from the C:Nratio given by Redfield et al. (1963). We did not calculatedCred based on the phosphate variation, because dCred variestoo widely with a small fluctuation of dCP.

The rightmost column in Tables 2a and 2b shows thecalculated dCred for each axis. In the density surfaces ofσθ ≥ 26.60 dCred seems to be constantly zero with a scatterof few µmol/kg, despite large potential uncertainty in bothdCN and dCTC in these density surfaces. This implies that themain source of negative dC values is the biological degrada-tion in the density surfaces, and diapycnal mixing is still notsignificant in these density surfaces. This also implies thatour calculations of CO(x) are appropriate and actual errors ofdC values due to the CO(x) error are rather small. If wecalculate CO(x) based on KH9605 Stn. 2 data, the resultingdCred varies by ±16 µmol/kg in the density surfaces of26.60 ≤ σθ < 26.90. In both axes a pulse of significantpositive dCred values are observed in the density surfacesaround σθ = 26.80. This seems to be because of the errone-ously high dCN values compared to dCP and dCTC inducedfrom the small difference in P, N and TC concentrationbetween Kuroshio and Oyashio in these density surfaces,rather than the influence of an actual influx of TC. If thedCred is due to the diapycnal mixing the surrounding densitysurface must also have significant dCred value.

Chemical Alternation in K/O Zone 693

In the density surfaces of σθ < 26.60 the dCred showssome scatter with a trend of negative dCred values. Thisimplies that the diapycnal mixing is also the source of dCvalues in these upper density surfaces.

There is some uncertainty about the C:N ratio in thebiological processes in Eq. (5). Many global-scale estima-tions (Takahashi et al., 1985; Anderson and Sarmiento, 1994)have found a C:N ratio similar to that of the original Redfieldratio of 6.6, but recent studies have shown that there may belarge local variations in the surface ocean with a generaltrend towards a high C:N ratio (e.g., Sambrotto et al., 1993;Karl et al., 1996; Chen et al., 1997). This may also be thecause of negative dCred values in the near-surface layers,although the extent is difficult to estimate.

4. Conclusive SummaryThroughout this study it has been exhibited that the

concentration of chemical properties in the K/O region showmore complex variations than that of salinity due to theprocesses in the source region such as biological processesand gas exchange. The low-PV low-salinity eddies seems toserve particular characters in the distributions of chemicalproperties in the K/O zone. The integrated transport ofchemical properties into the Kuroshio Extension axis, on theother hand, is shown to be almost equal to that of salinity inthe density surfaces with σθ greater than 26.90.

The change in chemical properties in each densitysurfaces of 26.60 ≤ σθ < 26.90 along the Kuroshio Extensionis explained only by isopycnal mixing of Kuroshio waterwith the Oyashio water and biological processes within theK/O zone. The influence of diapycnal mixing is estimated tobe small in the density surfaces greater than σθ = 26.60. Thisimplies that the amount of anthropogenic carbon injectedwithin the K/O region is also negligible in these densitylayers, because the density surfaces of σθ ≥ 26.60 are notoutcropped within the K/O zone and thus injection of an-thropogenic carbon in these density surfaces must requirediapycnal mixing. Many recent studies show that a significantamount of anthropogenic carbon is accumulated even in thedensity surface with σθ greater than 26.60 in the NPIW (e.g.,Chen, 1993; Tsunogai et al., 1993, 1995; Tokieda et al., 1996;Ono et al., 1998). Because the K/O zone is thought to be thearea of NPIW formation, observed small dCred values in-dicate that anthropogenic carbon in NPIW observed inprevious studies is injected into the NPIW source waterbefore the water is introduced into the K/O region. TheOkhotsk Sea and/or the area around the Kuril Islands may becandidates for such gas injection areas (Yasuda, 1997).

Strictly, there are some time lags between the sourcewater’s inflow into the K/O region and its outflow throughthe Kuroshio extension. Since our results are based onsynoptic surveys there may be some bias due to the lack ofinformation of the time lag. Further surveys with a smallertime interval are thus essential to study this area, which will

be carried out in future studies such as the Subarctic GyreExperiment (SAGE).

AcknowledgementsThe observational data used in this study were taken

during the collaborative study of Hokkaido University,Tohoku National Fisheries Research Institute and HokkaidoNational Fisheries Research Institute. We wish to expressour gratitude to all participants of WK9605 and HK9605cruises, the officers and crew of R/V Wakataka-Maru andHokko-Maru for their kind cooperation with the field work.We also thank all the co-workers of the above institutes fortheir valuable discussions and suggestions.

ReferencesAnderson, L. A. and J. L. Sarmiento (1994): Redfield ratios of

remineralization determined by nutrient data analysis. GlobalBiogeochem. Cycles, 8, 65–80.

Chen, C.-T. A. (1993): Anthropogenic CO2 distribution in the NorthPacific Ocean. J. Oceanogr., 49, 257–270.

Chen, C.-T. A., C.-M. Lin, B.-T. Huang and L.-F. Chang (1997):Stoichiometry of carbon, hydrogen, nitrogen, sulfur and oxygenin the particulate matter of the western North Pacific marginalseas. Mar. Chem., 54, 179–190.

Culberson, C., R. M. Pytkowitcz and J. E. Hawley (1970): Seawateralkalinity determination by the pH method. J. Mar. Res., 28,15–21.

Dickson, A. G. (1993): pH buffers for sea water media based onthe total hydrogen ion concentration scale. Deep-Sea Res., 40,107–118.

Dickson, A. G. and F. J. Millero (1987): A comparison of theequilibrium constants for the dissociation of carbonic acid inseawater media. Deep-Sea Res., 34, 1733–1743.

DOE (1994): Handbook of methods for the analysis of the variousparameters of the carbon dioxide system in sea water; ver. 2,edited by A. G. Dickson and C. Goyet, ORNL/CDIAC-74.

Hansson, I. (1973): A new set of constants for carbonic acid andboric acid in seawater. Deep-Sea Res., 20, 461–478.

Karl, D. M., J. R. Christian, J. E. Dore, D. V. Hebel, R. M. Letelier,L. M. Tupas and C. D. Winn (1996): Seasonal and interannualvariability in primary production and particle flux at StationALOHA. Deep-Sea Res. 2, 43, 539–568.

Levitus, S. and T. P. Boyer (1994): World Ocean Atlas 1994.NODC. NOAA Atlas NESDIS 4.

Maksimenko, N. A., T. Yamagata and K. Okuda (1997): Frontalconvection in Kuroshio and at the subarctic front. Oceanology,37, 295–300.

Nojiri, Y., J. Zeng, C. S. Wong and T. Kimoto (1997): Ship-of-Opportunity measurement of pCO2 in the Northern Pacificwith complete seasonal coverage. p. 87–90. In BiogeochemicalProcesses in the North Pacific: Proceedings of the Interna-tional Marine Science Symposium, ed. by S. Tsunogai, JapanMarine Science Foundation, Tokyo.

Ono, T., S. Watanabe, K. Okuda and M. Fukasawa (1998):Distribution of total carbonate and related properties in theNorth Pacific along 30°N. J. Geophys. Res. (in press).

Redfield, A. C., B. H. Ketchum and F. A. Richards (1963): The

694 T. Ono et al.

influence of organisms on the composition of seawater. p. 26–77. In The Sea, ed. by M. H. Hill, Wiley-Interscience, NewYork.

Rogachev, K. A., P. Y. Tishchenko, G. Y. Pavlova, A. S. Bychkov,E. C. Carmack, C. S. Wong and G. I. Yuasov (1996): Theinfluence of fresh-core rings on chemical concentrations (CO2,PO4, O2, alkalinity, and pH) in the western subarctic PacificOcean. J. Geophys. Res., 101, 999–1010.

Sambrotto, R. N., G. Savidge, C. Robinson, P. Boyd, T. Takahashi,D. M. Karl, C. Langdom, D. Chipman, J. Marra and L.Codispoti (1993): Elevated consumption of carbon relative tonitrogen in the surface ocean. Nature, 363, 248–250.

Shitashima, K., S. Kraines, D. Tsushi, I. Asanuma and T. Ono(1996): On the measurement of total carbonate dissolved inthe ocean. Abstracts of Spring Meeting of the OceanograpySociety of Japan, Tokyo, p. 269.

Takahashi, T., W. S. Broecker and S. Langer (1985): Redfieldratio based on chemical data from isopycnal surfaces. J.Geophys. Res., 90, 6907–6924.

Talley, L. D. (1988): Potential vorticity distribution in the NorthPacific. J. Phys. Oceanogr., 18, 89–106.

Talley, L. D. (1993): Distribution and formation of North PacificIntermediate Water. J. Phys. Oceanogr., 23, 517–537.

Talley, L. D., Y. Nagata, M. Fujimura, T. Iwao, T. Kono, D.Inagake, M. Hirai and K. Okuda (1995): North Pacific Inter-mediate Water in the Kuroshio/Oyashio Mixed Water Region.J. Phys. Oceanogr., 25, 473–501.

TNFRI (1996). 100 m Isotherms in the Tohoku Offshore Area,1996 May. Current status report of fisheries environments inthe Tohoku Offshore Area, 1997, June.

Tokieda, T., S. Watanabe and S. Tsunogai (1996): Chlorofluoro-carbons in the western North Pacific in 1993 and formation ofNorth Pacific Intermediate Water. J. Oceanogr., 52, 475–490.

Tsunogai, S. (1997): Why is the North Pacific absorbing muchanthropogenic CO2? p. 1–11. In Biogeochemical Processes inthe North Pacific: Proceedings of the International MarineScience Symposium, ed. by S. Tsunogai, Japan Marine ScienceFoundation, Tokyo.

Tsunogai, S., T. Ono and S. Watanabe (1993): Increase in totalcarbonate in the western North Pacific water and a hypothesison the missing sink of anthropogenic carbon. J. Oceanogr., 49,305–315.

Tsunogai, S., S. Watanabe, M. Honda and T. Aramaki (1995):North Pacific Intermediate Water studied chiefly with radio-carbon. J. Oceanogr., 51, 519–536.

Yasuda, I. (1997): The origin of the North Pacific Intermediatewater. J. Geophys. Res., 102, 893–909.

Yasuda, I., K. Okuda and M. Hirai (1992): Evolution of a Kuroshiowarm-core ring; variability of the hydrographic structure.Deep-Sea Res., 39, S131–S161.

Yasuda, I., K. Okuda and Y. Shimizu (1996): Distribution andmodification of North Pacific Intermediate Water in theKuroshio-Oyashio interfrontal zone. J. Phys. Oceanogr., 26,448–465.