INTRODUCTION
Dissolved inorganic carbon (DIC) is the largest exchangeable pool of carbon in the ocean and is biogeochemically important because it is directly linked to the organic carbon pool through photosynthesis and respiration. The variability of DIC in the ocean is controlled not only by anthropogenic CO2 invasion (Winn et al. Reference Winn, Li, Mackenzie and Karl1998; Tsunogai Reference Tsunogai2000) but also by changes in ocean circulation and biological activities (Tsurushima et al. Reference Tsurushima, Nojiri, Imai and Watanabe2002; Wakita et al. Reference Wakita, Watanabe, Murata, Tsurushima and Handa2010; Gruber Reference Gruber2011). Radiocarbon (14C) measurement of DIC has been a powerful tool to study carbon cycling (Broecker et al. Reference Broecker, Sutherland, Smethie, Peng and Ostlund1995; Levin and Hesshaimer Reference Levin and Hesshaimer2000), and circulation and transport of water masses in the ocean (Broecker et al. Reference Broecker, Peng, Ostlund and Stuiver1985; Key Reference Key1996; Druffel et al. Reference Druffel, Bauer, Griffin, Beaupré and Hwang2008; Kumamoto et al. Reference Kumamoto, Murata, Kawano, Watanabe and Fukasawa2013). In the last few decades, bomb 14C produced in the late 1950s and early 1960s entered the ocean through air–sea gas exchange and has provided an ideal tracer for short time scale (decades) water transport and circulation studies in the ocean (Stuiver and Östlund Reference Stuiver and Östlund1980; Stuiver et al. Reference Stuiver, Quay and Östlund1983; Druffel Reference Druffel1987; Broecker et al. Reference Broecker, Sutherland, Smethie, Peng and Ostlund1995). During the Geochemical Ocean Section Study (GEOSECS) in the 1970s, the World Ocean Circulation Experiment (WOCE) in the 1990s and the ongoing Climate Variability and Predictability Program (CLIVAR), a large number of DIC samples were collected from the world oceans and Δ14C measurements were reported (Östlund and Stuiver Reference Östlund and Stuiver1980; Key Reference Key1996; Stuiver et al. Reference Stuiver, Östlund, Key and Reimer1996; McNichol et al. Reference McNichol, Schneider, von Reden, Gagnon, Elder, NOSAMS, Key and Quay2000). These measurements have provided great insight into our understanding of the oceanic carbon cycle and ocean circulation processes, as linked to climate variability (examples are Stuiver et al. Reference Stuiver, Quay and Östlund1983; Broecker et al. Reference Broecker, Sutherland, Smethie, Peng and Ostlund1995; Key et al. Reference Key, Quay, Jones, McNichol, Reden and Schneider1996; Druffel et al. Reference Druffel, Bauer, Griffin, Beaupré and Hwang2008; Kumamoto et al. Reference Kumamoto, Murata, Kawano, Watanabe and Fukasawa2013; Sarnthein et al. Reference Sarnthein, Schneider and Grootes2013).
The northwest North Pacific (NP) is an important, highly dynamic region (Figure 1) where the Kuroshio Current (KC) and Oyashio Current (OC) mixes in the inter-frontal zone off the east coast of Japan, and the Kuroshio Extension (KE) flows eastward into the North Central Pacific (NCP) (Yasuda et al. Reference Yasuda, Okuda and Shimizu1996; Talley Reference Talley1997; Qiu and Chen Reference Qiu and Chen2011). The NP Intermediate Water (NPIW), which is formed in the mixed water region between the KE and Oyashio front (Talley Reference Talley1993), plays an important role, not only for its impacts on the global ocean circulation and regional climate variability (Yasuda et al. Reference Yasuda, Okuda and Shimizu1996; Qiu and Chen Reference Qiu and Chen2011; Wu et al. Reference Wu, Cai, Zhang, Nakamura, Timmermann, Joyce, McPhaden, Alexander, Qiu, Visbeck, Chang and Giese2012; Hu et al. Reference Hu, Wu, Cai, Gupta, Ganachaud, Qiu, Gordon, Lin, Chen, Hu, Wang, Wang, Sprintall, Qu, Kashino, Wang and Kessler2015), but also as a major sink for anthropogenic carbon dioxide in the northwest NP (Tsunogai et al. Reference Tsunogai, Ono and Watanabe1993). In the last two decades, many studies were conducted using chemical tracers such as 14C and chlorofluorocarbons (CFCs) to investigate the spatial and temporal variability of NPIW and its role influencing regional climate (Broecker et al. Reference Broecker, Spencer and Craig1982; Tsunogai et al. Reference Tsunogai, Watanbe, Honda and Aramaki1995; Tokieda et al. Reference Tokieda, Watanabe and Tsunogai1996; Takatani et al. Reference Takatani, Sasano, Nakano, Midorikawa and Ishii2012; Kumamoto et al. Reference Kumamoto, Murata, Kawano, Watanabe and Fukasawa2013). In a recent study, Kumamoto et al. (Reference Kumamoto, Murata, Kawano, Watanabe and Fukasawa2013) reviewed the decadal changes in bomb-produced 14C measured in DIC in the NP from the 1990s to the 2000s. By comparing the differences in decadal changes of vertical profiles of Δ14C-DIC, they found that the inventory of bomb 14C in the northwestern subtropical region decreased more significantly than that in the southern subtropical region due mostly to the 14C decrease in the upper thermocline. The decadal variations in bomb 14C in the NP indicate that the turnover time of thermocline circulation in the northwestern subtropical region is faster than that in the southern subtropical region as related to the transport of NPIW (Kumamoto et al. Reference Kumamoto, Murata, Kawano, Watanabe and Fukasawa2013). Using seven years (2002–2008) of surface DIC concentration data, Yasunaka et al. (Reference Yasunaka, Nojiri, Nakaoka, Ono, Mukai and Usui2014) reported that the observed surface DIC seesaw variations in the NP was related to the Pacific Decadal Oscillation (PDO).
Figure 1 Map showing the study site and the six sampling stations (marked with black circles) in the northwestern North Pacific in April–May 2015. For comparison, we also plot three stations studied previously. Stations P10-76 and P10-78, which are close to our B8 and B9 stations, were from the WOCE P10 line collected in 1993 (Key et al. Reference Key, Quay, Jones, McNichol, Reden and Schneider1996), and station P10N-120, which is close to our B2 station, was collected in 2005 during a repeating cruise (MR05-02) to P10 line by Japanese scientists referred to as P10N line (Kumamoto et al. Reference Kumamoto, Murata, Kawano, Watanabe and Fukasawa2013).
It has been 10 years since the collection of water samples for 14C measurement in the northwestern NP reported by Kumamoto et al (Reference Kumamoto, Murata, Kawano, Watanabe and Fukasawa2013). In this paper, we present ∆14C-DIC results of samples collected from six deep stations in the KE region in the northwestern NP in 2015. We observed large spatial variations in Δ14C-DIC values in the upper 1500 m among the stations. These variations in Δ14C-DIC profiles thus could provide useful information for our understanding of the water mass movement and mixing dynamics in the KE region.
METHODS
Oceanographic Setting
Our study sites in the western NP covered an area from 25ºN to 37ºN, and 134ºE to 152ºE (Figure 1). This is the region that is influenced largely by two major oceanic currents, the Kuroshio and Oyashio currents, which carry NPIW as part of the Pacific western boundary currents (Yasuda et al. Reference Yasuda, Okuda and Shimizu1996; Qiu and Chen Reference Qiu and Chen2011; Hu et al. Reference Hu, Wu, Cai, Gupta, Ganachaud, Qiu, Gordon, Lin, Chen, Hu, Wang, Wang, Sprintall, Qu, Kashino, Wang and Kessler2015). The northeastward-flowing Kuroshio Current separates from the coast of Japan at about 34ºN, 140ºE and flows eastward as the Kuroshio Extension into the North Central Pacific (NCP) (Yasuda et al. Reference Yasuda, Okuda and Shimizu1996; Qiu and Chen Reference Qiu and Chen2011). The Oyashio is a southward-flowing current along the east coast of Japan carrying cold and fresh subarctic water. At about 37ºN, the Oyashio front meets the KE and forms the Kuroshio-Oyashio interfrontal zone, where the subarctic water mass mixes with the relatively warm and saline KE water, and flows eastward (Yasuda et al. Reference Yasuda, Okuda and Shimizu1996; Yasuda Reference Yasuda2004; Hu et al. Reference Hu, Wu, Cai, Gupta, Ganachaud, Qiu, Gordon, Lin, Chen, Hu, Wang, Wang, Sprintall, Qu, Kashino, Wang and Kessler2015). The newly formed NPIW is characterized by a salinity minimum zone in the density range of 26.6–27.4 due to the along isopycnal mixing between the Kuroshio and Oyashio waters, and consists of ~55% Kuroshio and 45% Oyashio waters (Talley et al. Reference Talley, Nagata, Fujimura, Iwao, Kono, Inagake, Hirai and Okuda1995; Talley Reference Talley1997; Yasuda et al. Reference Yasuda, Okuda and Shimizu1996). The new NPIW is transported eastward by the KE as a low salinity tongue, with eventual mixing into the intermediate layers of the NP subtropical gyre (Yasuda et al. Reference Yasuda, Okuda and Shimizu1996; Qiu and Chen Reference Qiu and Chen2011). Long-term satellite data analyses have shown that the KE has undergone decadal oscillations between the elongated state and the contracted state, which affects both oceanic circulation and climate variations in the western NP (Qiu and Chen Reference Qiu and Chen2011; Hu et al. Reference Hu, Wu, Cai, Gupta, Ganachaud, Qiu, Gordon, Lin, Chen, Hu, Wang, Wang, Sprintall, Qu, Kashino, Wang and Kessler2015). Recent studies have demonstrated that sub-mesoscale dynamics, such as eddies in the KE, undergo seasonal cycles and play an important role regulating the surface currents in the region (Ma et al. Reference Ma, Zhao, Chang, Liu, Montuoro, Small, bryan, Greatbatch, Brandt, Wu, Lin and Wu2016; Rocha et al. Reference Rocha, Gille, Chereskin and Menemenlis2016). Our study sites fall within the Kuroshio-Oyashio interfrontal zone of the KE (Figure 1).
Study Sites and Sample Collection
Water samples were collected at six stations on the R/V Dongfanghong-2 a cruise from March 31 to May 7 in 2015 (Table 1, Figure 1). The other three stations: P10-76 and P10-78 (red triangles) and P10N-120 (green square) are from previous studies (Key Reference Key1996; Kumamoto et al. Reference Kumamoto, Murata, Kawano, Watanabe and Fukasawa2013). The deepest bottom depth in the KE region is 6200 m, however we collected samples down to only 4000 m (due to an unexpected technical problem). Water samples were collected using 12-L Niskin bottles deployed on a rosette with CTD. For DIC water collection, we used the protocol described in McNichol and Jones (Reference McNichol and Jones1991). Water was collected (after overflowing ~100 mL) in prebaked 100-mL glass bottles, with fine ground-glass stoppers, using precleaned silicone tubing connected directly to the Niskin bottle. After adding 50 μL saturated HgCl2 solution, the bottles were capped tightly with grease-coated, ground-glass stoppers and secured with rubber bands to make a gas-tight seal to avoid any CO2 exchange with atmosphere. All samples were kept in the dark at room temperature.
Table 1 Description of the sampling site in the northwestern North Pacific
KE: Kuroshio Extension.
DIC Extraction and Concentration Analysis
After the cruise, DIC samples were processed in the laboratory within one month. DIC was extracted as gaseous CO2 using our modified method (Ge et al. Reference Ge, Wang, Zhang, Luo and Xue2016) based on McNichol et al. (Reference McNichol, Jones, Hutton and Gagnon1994). Briefly, in a N2 filled glove-bag, a 50 mL water sample was injected into a pre-evacuated 100-mL borosilicate glass bottle with ground-glass-joint stripping probes (precombusted at 550ºC for 4 hr). After injecting 1.0 mL high purity 85% H3PO4 using a glass syringe and stainless steel needle, the glass bottle with acidified water sample was placed in a hot water bath (70ºC) for 30 min and shaken by hand several times. At pH≤2, all forms of DIC (carbonate, bicarbonate, and CO2) dissolved in water will become CO2. It is calculated that at 70ºC under vacuum, the solubility of CO2 in water is <6.0×10–5 μmol/L, so it is expected that all CO2 dissolved in water has escaped into the probe headspace. The glass bottle was removed from the hot water bath and cooled for 5 min and connected to the vacuum line. All CO2 generated was frozen into a liquid nitrogen trap. After the volume was measured manometrically, the CO2 was flame-sealed inside 6-mm OD Pyrex tubes for 14C and 13C analyses. The extraction efficiency of DIC with the method was >96% (Ge et al. Reference Ge, Wang, Zhang, Luo and Xue2016).
Concentrations of DIC were measured using a Shimadzu TOC-L analyzer equipped with an ASI-V auto-sampler, using the total inorganic carbon (IC) mode. The concentration of DIC was calibrated using a 5-point calibration curve prepared from reagent grade sodium carbonate and sodium bicarbonate dissolved in DIC-free Milli-Q water as for the IC standard. The instrument blank and DIC values were checked against DIC reference materials (CRMs provided by Dr. A. Dickson at Scripps Institution of Oceanography, University of California San Diego). Total blanks associated with DIC measurements were less than 3.0 μM, which is<0.15% of seawater DIC concentrations, and the analytic precision of triplicate injections was <2%.
Isotopic Measurements
Both δ13C and Δ14C of DIC samples were analyzed at the National Ocean Sciences Accelerator Mass Spectrometry (NOSAMS) facility at Woods Hole Oceanographic Institution (WHOI). A small split of CO2 was measured for δ13C using a VG Isotope Ratio Mass Spectrometer (IRMS) and the remaining CO2 was graphitized for Δ14C analysis using AMS. Values of δ13C are reported in ‰ relative to the VPDB standard and the 14C measurements were reported as modern fraction (McNichol et al. Reference McNichol, Jones, Hutton and Gagnon1994). The conventional radiocarbon ages (yr BP) were calculated using the Libby half-life as ascribed by Stuiver and Polach (Reference Stuiver and Polach1977). The total uncertainty of Δ14C-DIC and δ13C-DIC analyses are 6‰ and 0.1‰ or better, respectively, tested with a DIC standard (Ge et al. Reference Ge, Wang, Zhang, Luo and Xue2016).
RESULTS
The hydrographic data, DIC concentrations and isotopic results measured for 64 samples collected from the six stations are summarized in Table 2.
Table 2 Concentration and carbon isotopic (δ13C and Δ14C) compositions of DIC measurements in the water samples collected from the western North Pacific.
Hydrographic Profiles
Figure 2 shows the temperature, salinity, and density profiles for the six stations. Temperature (T) decreased rapidly from the surface to about 1500 m depth and then remained constant for all stations (Figure 2a). The largest differences in the T profiles appear in the upper 600 m depth, where Sta K2 had the highest T (24.40ºC) and Sta B2 had the lowest T (14.72ºC) in the surface layer (5 m). Water temperature profiles for stations A4, A8, B8, and B9 showed less variation and decreased in parallel from the surface (17.5–20.7ºC) to about 2.5ºC at 1500 m and then remained the same below 1500 m (Figure 2a). Salinity decreased from the surface initially and then increased with depth to about 2500 m. Below 2500 m, the salinity for all stations were the same and remained relatively constant. Like T, the largest variations in S profiles are also seen in the upper 500 m depth. Sta B2 had the lowest salinity (34.490) in the surface and decreased to 33.662 by 250 m, then increased to similar values as the other five stations at 2500 m. The S profiles for stations K2, A4, A8, B8, and B9 showed less variations in the surface water (5 m) (34.877–34.982) and Sta K2 had the highest S value (34.982) among all stations in the surface. The potential density profiles also clearly show the distinct water masses in the upper 1000 m depth (Figure 2c) with large differences among the stations. Sta B2 had the highest potential density in the upper 1000 m depth, and Sta K2 had the lowest water density values in the upper 200 m, respectively.
Figure 2 Temperature, salinity and potential density profiles for the six stations in the northwestern North Pacific sampled in April–May 2015. The insert figure shows the profile in the upper 1000 m depth.
DIC and Isotopic Distributions
Profiles of DIC concentration and their isotopic values for the six stations are shown in Figure 3. For the Δ14C-DIC profiles (Figure 3b), we also plot three stations studied earlier for comparison (Figure 1). Stations P10-76 and P10-78 that are close to our B8 and B9 stations were from the WOCE P10 line collected in 1993 (data provided by Dr. Robert Key at Princeton University). Station P10N-120 that is close to our B2 station was collected in 2005 during a repeat cruise (MR05-02) to the P10 line by Japanese scientists referred to as the P10N line (Kumamoto et al. Reference Kumamoto, Murata, Kawano, Watanabe and Fukasawa2013). These stations in the same region were sampled ~10 years apart, so they provide data comparison for the last 22 years (1993–2015).
Figure 3 Depth profiles of DIC, Δ14C-DIC and δ13C-DIC measured for the six stations in the northwestern North Pacific sampled in April–May 2015. In Figure 3b, the depth profiles of Δ14C-DIC for stations P10-76, P10-78, and P10N-120 are plotted for comparison (see Discussion). Note: the depth scale below 1500 m has been reduced.
Concentrations of DIC ranged from 1798 to 2141 μmol/kg (Figure 3a, Table 2). In general, concentrations of DIC showed similar values in the surface water (5 m) and increased with depth even though the trend exhibited some variations among the stations (Figure 3a). Concentrations of DIC showed less variation in the upper 200 m and all increased at depths below 200 m. The most variability in DIC concentration at all stations was observed between 400–1000 m depths. For the four deep stations (A4, A8, B2, and B8), concentrations of DIC for A8 and B8 remained relatively constant from 1500–3500 m; decreased in the middle depth (2000 m) and then increased and remained constant at 3000–4000 m for station A4; and slightly decreased from 2000–4000 m depth at station B2 (Figure 3a). Station A8 in the middle of the KE had the highest DIC concentrations (2130–2141 μmol/kg) from 1500–3500 m.
Values of Δ14C-DIC in the surface water (5–50 m) ranged from 20‰ to 52‰ for the six stations and all decreased with depth to –200‰ to –220‰ at about 1500 m, then remained relatively constant in the deeper waters (Figure 3b). As for T and S profiles, large variations in Δ14C-DIC were seen in the upper 1000 m among the stations. Values of Δ14C-DIC from Sta A4 decreased to the lowest values (–233‰) by 500 m, a 278‰ drop in Δ14C values in just 300 m of the water column. Below 500 m, the Δ14C-DIC values for Sta A4 remained constant down to 4000 m. Values of Δ14C-DIC at Sta B2 also showed a rapid decrease from 32‰ in the surface to –234‰ at 1000 m depth, then remained similar to those at Sta A4 by 4000 m depth (Figure 3b). Δ14C-DIC for Sta B8 decreased rapidly, first from the surface value (49‰) to –99‰ at 500 m, and then to –197‰ at 1500 m. The Δ14C-DIC profiles for the stations K2, A8, and B9 exhibited a similar trend, showing relatively constant values in the upper 500–600 m and then decreased to 1500 m (1000 m for K2). At 500-m depth, there were much lower ∆14C values at A4 compared with those at A8, B9, and K2 (Figure 3b). The differences of the Δ14C-DIC profiles between our stations and those from the three earlier stations are presented in the Discussion section.
The δ13C-DIC values from the six stations ranged from 0.46‰ to –1.87‰, and the highest variation appeared in the depths from 100–500 m (Figure 3c). Values of δ13C-DIC showed a low peak for stations B8 and B9 at 200 m depth, and K2 at ~500 m depth. Below 500 m depth, the δ13C-DIC values were generally constant for all stations.
DISCUSSION
Variability of Δ14C-DIC in the Kuroshio Extension
The KE is a highly dynamic region where the Kuroshio and Oyashio Currents meet and influence the hydrology and chemical properties of the resulting water mass (Tsunogai et al. Reference Tsunogai, Watanbe, Honda and Aramaki1995; Tsurushima et al. Reference Tsurushima, Nojiri, Imai and Watanabe2002). The DIC concentration and isotopic profiles for the six stations (Figure 3) indicate the influence of different water masses in the region, especially in the upper 1000 m of the water column. In Figure 4, we plot T versus salinity with DIC concentrations and Δ14C-DIC values (as the colors of the points) associated with the potential density anomaly. It can be clearly seen that the distributions of both DIC concentrations and Δ14C-DIC were associated with different density water masses in the region (groups of circled points). High DIC concentrations and low Δ14C-DIC values were present in denser waters (σ0 > 27.0), while lower DIC concentrations and high Δ14C-DIC values were present in lower density water masses (σ0<25.5) (Figure 4). The denser water masses likely originated from the subarctic gyre, which circulates with the NPIW and has low T and S values. Based on the field observations, Yasuda (Reference Yasuda2004) reported that NPIW circulation in the subarctic gyre is related to diapycnal-meridional overturning, generated around the Okhotsk Sea due to tidal-induced diapycnal mixing and dense shelf water formation accompanied by sea-ice formation; and transport through the south-flowing Oyashio along the western boundary from the subarctic to KE region. The lower density water masses with high T and S, however, were transported by the northeast-flowing Kuroshio Current and mixed with Oyashio water in the KE region (Figure 2).
Figure 4 Plot T versus salinity with (a) DIC concentration and (b) Δ14C-DIC values (indicated as colors) for all samples measured for the six stations in the northwestern North Pacific. Lines are constant potential density. The circular areas represent different water masses as (A) lighter water mass in the upper 300 m depth with lower DIC and higher Δ14C-DIC; (B) mixed upper water in 300–500 m depth; (C) mixed intermediate water in 500–800 m depth; and (D) denser NP deep water in >1000 m depth. The one outlier at the right-top is the surface water of Station K2. (Colors refer to online version.)
Examination of the ∆14C-DIC profiles reveals that Sta K2, which is on the path of the north-flowing Kuroshio Current, is a region that transports water with relatively higher Δ14C-DIC values in the upper 500 m (38–60‰). These Δ14C-DIC values shows that bomb 14C was well mixed down to 500 m depth in the Kuroshio Current (Figure 3b). For stations A8 and B9 in the KE region, the Δ14C-DIC profiles also showed similar high bomb 14C signals that were mixed down to ~500 m depth, likely influenced by the Kuroshio water mass. In comparison, for stations A4 and B8, the well-mixed high bomb 14C signal was observed in the upper 150–250 m only and the Δ14C-DIC values decreased rapidly with depth (Figure 3b). As discussed above, the Δ14C-DIC profiles for stations B2, A4 and B8 appear to be influenced by the Oyashio Current, which carries subarctic NPIW southward and mixes with the Kuroshio water in the region around 35–37ºN and 143ºE east (Yasuda et al. Reference Yasuda, Okuda and Shimizu1996; Qiu and Chen Reference Qiu and Chen2011). In its path, this subarctic water mass influences the T and S profiles of Sta B2 in the upper 600 m depth as being well demonstrated in Figure 2. And this subarctic water mass also carries higher DIC concentrations affecting stations B2 and A4 in the upper 700 m depth as well (Figure 3a).
To further demonstrate the influence of different water masses on the chemical properties in KE region, Figure 5 compares the cross sectional distributions for salinity, DIC concentration and Δ14C-DIC for the four stations B2, A4, B8, and B9 as a line transect across the KE from 37º00′N to 29º51′N. The regional influences of the two water masses carried by the Kuroshio and Oyashio currents can be seen more clearly in Figure 5. It appears likely that the subarctic water with low salinity, high DIC concentrations and low Δ14C-DIC values carried by the Oyashio Current intruded southward to about 34ºN in the upper 500 m at Sta A4 and mixed with the Kuroshio water to form the KE water mass. However, we are not able to determine if this southward intrusion of the Oyashio water reflects a seasonal change or a decadal oscillation of the NPIW.
Figure 5 Comparison of cross section distributions of salinity, DIC concentration and ∆14C-DIC values for stations B9, B8, A4, and B2 in the northwestern North Pacific.
It is possible that the large Δ14C-DIC variability we observed between stations A4, B2 and others could be largely influenced by local water circulation, vertical mixing processes and/or mesoscale eddies in the KE region. Physical oceanographic studies have shown that the formation of the KE along the boundary between the NP subtropical and subarctic gyres is hydrodynamically unstable, thus allowing vigorous ocean eddies to form in the KE region (Qiu and Chen Reference Qiu and Chen2011; Waterman et al. Reference Waterman, Hogg and Jayne2011; Ma et al. Reference Ma, Zhao, Chang, Liu, Montuoro, Small, bryan, Greatbatch, Brandt, Wu, Lin and Wu2016). In their recent study, Rocha et al. (Reference Rocha, Gille, Chereskin and Menemenlis2016) reported that the seasonal change of upper ocean stratification in the KE also modulates sub-mesoscale (10–100 km) inertia-gravity waves and turbulence. These inertia-gravity waves and turbulence could affect the chemical profiles in the region. Qiu and Chen (Reference Qiu and Chen2011) also reported that the KE is an unstable current that can transport northern Oyashio-origin water southward through mesoscale eddies, causing the fresher Oyashio water to be “diffused” southward across the KE. The significant, low Δ14C-DIC values we observed at stations B2 and A4 in the upper 200–1000 m depth could reflect this transport south via mesoscale eddies. A recent high-resolution coupled atmosphere-ocean model has revealed that local ocean mesoscale eddies and atmospheric feedback is fundamental to the dynamics governing the KE variability (Ma et al. Reference Ma, Zhao, Chang, Liu, Montuoro, Small, bryan, Greatbatch, Brandt, Wu, Lin and Wu2016). If this is the case, we could expect that the southward intrusion of the Oyashio Current that carries fresh, cold, nutrient-rich, high DIC subarctic NPIW could have a significant influence on the chemical and biological processes in the KE region. Nutrient-rich NPIW could enhance primary production in the region, and thus increase the net flux of atmospheric CO2 into the surface ocean. The cold, subarctic water mixing with warm Kuroshio water could also reduce the heat transfer to the atmosphere, and thus affect the regional climate (Wu et al. Reference Wu, Cai, Zhang, Nakamura, Timmermann, Joyce, McPhaden, Alexander, Qiu, Visbeck, Chang and Giese2012; Hu et al. Reference Hu, Wu, Cai, Gupta, Ganachaud, Qiu, Gordon, Lin, Chen, Hu, Wang, Wang, Sprintall, Qu, Kashino, Wang and Kessler2015; Ma et al. Reference Ma, Zhao, Chang, Liu, Montuoro, Small, bryan, Greatbatch, Brandt, Wu, Lin and Wu2016).
The depth profiles of δ13C-DIC for the stations show large variations, mainly in the upper 500 m depth (Figure 3c). The relatively higher δ13C-DIC values in the surface waters (0–50 m) reflect the influence of photosynthesis, which preferentially fix light 12C into OC and thus leaves 13C enriched DIC in the surface waters. On the other hand, sinking and remineralization of OC below the euphotic zone releases light CO2, and may have caused the observed decrease of δ13C-DIC values at 200–500 m depths at stations B9 and B8, respectively (Figure 3c). During our sampling in May–June 2015, we measured very low dissolved nitrogen (DN) concentrations (<4 μM in the upper 100 m; Ding et al. Reference Ding, Ge, Gao, Zhang and Wang2017) in the surface waters for the stations, suggesting the primary production was active in the region. This supports our DIC and Δ14C-DIC profiles discussed above.
Comparison with Earlier Δ14C-DIC Results
During the WOCE Program in the early 1990s, thousands of DIC samples were collected from the world oceans and their Δ14C values reported (Key Reference Key1996; Stuiver et al. Reference Stuiver, Östlund, Key and Reimer1996; McNichol et al. Reference McNichol, Schneider, von Reden, Gagnon, Elder, NOSAMS, Key and Quay2000). Δ14C-DIC measurements have provided considerable insight to the understanding of the oceanic carbon cycle and ocean circulation processes, as linked to climate variability. Here, we compare our Δ14C-DIC results with those from three earlier stations in the KE region. As shown in Figure 3b, Stations P10–76 and P10–78 sampled in 1993 had much higher Δ14C-DIC values (~130‰) in the surface and decreased rapidly down to 1500 m depth. Below 700–800 m depth, the Δ14C-DIC profiles of P10–76 and P10–78 had similar values and decreased in parallel as our B8, B9, and A8 stations, suggesting that the same water mass was carried by the KE in the last 20 years. The higher Δ14C-DIC values in the upper 700 m depth of stations P10–76 and P10–78 reflect the higher bomb–produced 14C present in the surface down to 600 m depth in 1993. The different values of Δ14C-DIC between the two P10 stations and B8 and B9 (80–90‰) in the surface was likely due to the decrease of the bomb Δ14C signal in the atmosphere in the last 20 years that has mixed into the upper ocean through air–sea exchange. As reported by Hua et al (Reference Hua, Barbetti and Rakowski2013), the global average Δ14C value in atmospheric CO2 has decreased from ~150‰ in 1990 to ~50‰ in 2010, a 5‰ drop annually over the past 20 years. This has lowered the Δ14C-DIC signature of the upper ocean, thus explaining the observed Δ14C-DIC differences in the surface between the two P10 stations and our sites (Figure 3b). However, the large difference of Δ14C-DIC (~150‰) between the two P10 stations and B8 in the depths of 200–500 m cannot be similarly explained. As can be seen, the Δ14C-DIC profile of Sta B8 in the upper 500 m was the same as the Sta P10N-120 sampled in 2005. As discussed above, we expect that the Δ14C-DIC profile in the upper 800 m depth at stations B2 and A4 were all influenced by the subarctic NPIW carried by the southward-flowing Oyashio Current during our sampling. The relatively lower Δ14C-DIC values measured in the 250–500-m depth at Sta B8, however, could be affected by the seasonal or annual variations of the local circulation, vertical mixing and mesoscale eddies (Qiu and Chen Reference Qiu and Chen2011; Waterman et al. Reference Waterman, Hogg and Jayne2011; Ma et al. Reference Ma, Zhao, Chang, Liu, Montuoro, Small, bryan, Greatbatch, Brandt, Wu, Lin and Wu2016) which upwelled some deep water with lower Δ14C-DIC to the upper depth at Sta B8. More seasonal studies are needed to better answer these questions.
Application of Keeling Plot on DIC Mixing in KE
Keeling plot analysis has been performed on DIC depth profiles to discern consistency with solution based two-component mixing model (Beaupré and Aluwihare Reference Beaupré and Aluwihare2010). As plotted in Figure 6, there is a good correlation between Δ14C-DIC and [DIC] –1 (r2=0.84, p<0.001) in the KE, suggesting that both parameters were mainly controlled by the hydrodynamic mixing processes in the KE region. When separating the water samples into three depths, it can been seen that some deep water with low Δ14C-DIC values were mixed into the upper water column (0–500 m). The slope (4.0×106) and intercept (–2026) values of the Keeling plot for the six stations in the KE are consistent with the values obtained in the South Ocean (5.64×106, –2670), eastern NP (4.29×106, –2040), and central NP (5.28×106, –2470) as reported by Beaupré and Aluwihare (Reference Beaupré and Aluwihare2010).
Figure 6 Keeling plot of Δ14C-DIC versus concentration of [DIC] –1 measured for the six deep stations in the Kuroshio Extension region in the northwest North Pacific. The line is a linear regression fit to all data points.
When they examined large data sets of concentrations of DIC versus Δ14C-DIC, Sarnthein et al. (Reference Sarnthein, Schneider and Grootes2013) obtained regression slopes of –0.79 and –1.49 for the South Pacific and North Pacific, respectively. Our value of –0.85 (plot not shown) is also comparable to their values. These similarities indicate that the distribution of Δ14C-DIC in the open ocean could be predicted by the solution mixing model and DIC can be used as a conservative tracer for water mass movement and water parcel homogenization (Beaupré and Aluwihare Reference Beaupré and Aluwihare2010). If we take an average Δ14C-DIC value of 50‰ for the Kuroshio water in the upper 500 m depth (Sta K2) and –220‰ for the NPIW of Oyashio (Figure 6), we calculated based on the mass balance that 55–58% Oyashio water and 42–45% of Kuroshio water could be mixed at the depth of 500 m to result in the observed Δ14C-DIC values (Figure 3b) at stations B2 and B8; 100% Oyashio water at Sta A4 and 96–100% Kuroshio water at stations A8 and B9, respectively. These mixing ratios are clearly restricted to a surface water scenario where it is affected by the present state of the bomb signals. The more natural mixing dynamics of Δ14C-DIC of water samples collected in >2000 m depth in the global oceans have been recently examined and referred to Sarnthein et al. (Reference Sarnthein, Schneider and Grootes2013).
SUMMARY
The large variability of Δ14C-DIC values that we observed in the KE in the northwestern NP revealed the complicated mixing dynamics of the different water masses that influence the chemical and isotopic distributions in the region. The results suggest that the low Δ14C-DIC and high DIC profiles at stations A4 and B2 were largely affected by the Oyashio Current front which carries the cold and fresh subarctic intermediate water into the KE. In contrast, the relatively higher Δ14C-DIC and low DIC profiles at stations A8, B9, and K2 were likely influenced by the northeast-flowing Kuroshio Current, which carries warm and high salinity water. The Δ14C-DIC and DIC values can be used as relatively conservative tracers of water mass movement and mixing in the ocean. The large variability in the observed Δ14C-DIC profiles could have also been caused by mesoscale eddies in the KE region. In any case, the intrusion of nutrient-rich subarctic water with low Δ14C-DIC values could have an important influence not only on the regional production and ecosystem, but also on carbon cycling in the northwestern NP. Further studies are required to determine the actual temporal variability of this system.
ACKNOWLEDGMENTS
We thank Drs. Lixin Wu and Xiaopei Lin for providing the cruise opportunity and the crews of the R/V Dongfanghong-2 and Dr. Lei Li for the help during sample collection. We thank Dr. Robert Key at Princeton University and Dr. Yuichiro Kumamoto at the Research Institute for Global Change, Japan Agency for Marine–Earth Science and Technology (JAMSTEC) for kindly providing the data for comparison. We also thank Dr. Ann McNichol and the staff and colleagues at the WHOI National Ocean Science Accelerator Mass Spectrometry (NOSAMS) facility for measurements of δ13C and Δ14C of the samples. We appreciate the thorough review and comments from Associate Editor P. Grootes and two reviewers. Financial support for this work was provided by China’s National Natural Science Foundation (Grants: 91428101, 41476057) and the Fundamental Research Funds for the Central Universities (Grant: 201762009).
