Vertical turbulent nitrate flux from direct measurements in the western subarctic and subtropical gyres of the North Pacific

Vertical turbulent nitrate fluxes were estimated in the western North Pacific from direct measurements of vertical turbulent mixing and vertically continuous nitrate profiles during the summer of 2008. We made three north–south transects that covered the area from the subarctic to the subtropics including a section along the Emperor Sea Mounts. Subsurface fluxes generally showed an increasing trend with increasing vertical gradient of nitrate from oligotrophic subtropical to non-oligotrophic subarctic waters. Enhanced fluxes [O(10−6) mmol m−2 s−1] due to elevated mixing [vertical diffusivity: O(10−5) m2 s−1] were observed, especially over the Emperor Sea Mounts. It is suggested that the internal tide generated by the topography enhanced the vertical mixing. In other subarctic areas, the fluxes were estimated as O(10−7) mmol m−2 s−1. The same order of fluxes was also found in the frontal area between the subarctic and subtropical gyres, the Kuroshio–Oyashio Transition Area. Enhancement of fluxes in the frontal area, including the Kuroshio Extension, was also observed at mid-depth regions, and their vertical divergence suggested nitrate transport from North Pacific Intermediate Water to lighter densities. In the frontal areas, the enhancement of turbulence is caused by the surface wind rather than the internal tide. In contrast, in the subtropical regions, subsurface fluxes were estimated as O(10−8) mmol m−2 s−1 owing to the small nitrate gradient even where diffusivity was enhanced. In these regions, enhancement of diffusivity, including that at mid-depths, corresponded to the elevation of the internal-tide dissipation, in addition to that of surface turbulence.


Introduction
Turbulent mixing is an essential process for the vertical transport of heat, salt, and substances, including nutrients, in the stratified ocean (e.g., Thorpe 2004). Such vertical nutrient transport also has an important role in phytoplankton production, taxonomic composition, and size structure in the euphotic layer (Chisholm 1992). Vertical turbulent nitrate flux is a key component of primary production (Eppley et al. 1979;Lewis et al. 1986;Carr et al. 1995). It is one of the main pathways for nutrient supply to the euphotic layer in regions where the surface water is permanently (e.g., subtropical oceans) or seasonally (e.g., subarctic oceans) deficient in nitrate. Nitrogen fixation is another source of nitrogen for new production in regions deficient in surface nitrate (e.g., Mouriño-Carballido et al. 2011;Horii et al. 2018). Evaluation of the vertical turbulent nitrate flux based on direct observations allows us to investigate the surface nitrogen path quantitatively. Moreover, observations of vertical turbulent fluxes into mid-depth regions also provide information on nitrate transport in deeper waters that is also important for understanding biogeochemical cycles.
Estimates of the vertical turbulent nitrate flux are obtained in practice by multiplying the vertical nitrate gradient by the vertical eddy diffusivity (e.g., Lewis et al. 1986). Although several studies of turbulence with basinscale coverage have advanced knowledge about turbulence in the North Pacific (Moum and Osborn 1986;Nagasawa et al. 2007;Mori et al. 2008;Fernández-Castro et al. 2014;Itoh et al. 2020), concurrent measurements of the nitrate profile and vertical diffusivity are rare except over the western boundary region (e.g., Kaneko et al. 2013;Nagai et al. 2019;Tanaka et al. 2019) and eastern boundary region (Hayward 1987;Fernández-Castro et al. 2015). Therefore, basin-scale variability of vertical nitrate flux in the region has yet to be clarified, especially in the western North Pacific. From the perspective of nutrient circulation, the western North Pacific is important because the region has a distinctive circulation, whereby North Pacific Intermediate Water originating from the Okhotsk Sea (Yasuda 1997) has been identified as affecting nutrient distribution from the surface to mid-depths, which differs from the influence of Subantarctic Mode Water in other oceans (Sarmiento et al. 2004).
In the present study, we focus on the basin-scale variability of vertical nitrate flux in the western North Pacific. The western North Pacific is divided by wind-driven circulation into the subarctic gyre, the subtropical gyre, and the tropical region. The subarctic gyre is characterized by an intense maximum of the vertical gradient of nutrients (i.e., nitracline) and pycnocline below the surface waters (from the surface to dozens of meters in depth) owing to upwelling caused by the surface wind-driven circulation, whereas the subtropical gyre has deeper nitracline and pycnocline caused by downwelling. Thus, it is expected that active vertical nitrate transport into the euphotic zone occurs in the subarctic gyre. In addition to such gyre-scale contrast, the local variability in fluxes may be modified by turbulence in the subsurface pycnocline (including the seasonal thermocline, at depths of dozens to several hundreds of meters) caused by local topographic and atmospheric influences, and the relative vertical position of the nitracline. Subsurface turbulence in the open ocean is generally caused by internal waves generated from internal tides around characteristic topography, including the Emperor Sea Mount Chain (Niwa and Hibiya 2001;Nagasawa et al. 2007), as well as by the surface wind. Thus, the Emperor Sea Mount Chain (ESM) might also have a specific ecological character; if the level of the nitrate flux differs from that in surrounding basin areas.
Another subarea of the western North Pacific between the southern bound of the subarctic gyre and the northern bound of the subtropical gyre is expected to have the characteristics of an ecotone. The bounds are independently recognized as the subarctic front and the Kuroshio Extension (KE), and the area between the bounds is known as the Kuroshio-Oyashio Transition Area (KOTA; Yasuda 2003). The KOTA contains many mesoscale eddies and fronts and has a complex distribution of water masses. Previous observations conducted in the western boundary region have indicated that the nitrate flux is enhanced in the frontal areas (Kaneko et al. 2013;Nagai et al. 2019;Tanaka et al. 2019). Strong currents accompanied by intense fronts can affect the intensity of turbulence by modulating the propagation and breaking of internal waves (e.g., Whalen et al. 2018) and through direct energy supply (e.g., D'Asaro et al. 2011). Thus, turbulence enhancement and subsequent elevated vertical fluxes may occur in regions with complicated frontal structures such as the KOTA. The nature of turbulence is such that enhancement would be expected to occur intermittently and unevenly in the vertical direction rather than continuously and over a large depth range. This generally makes the estimation of the level of turbulence and fluxes difficult; however, fluxes are rarely measured in the oceanic part of the KOTA, so it is worthwhile conducting and accumulating estimates based on field observations. With respect to nitrate circulation, particularly at middepths, the North Pacific Intermediate Water plays a key role in nutrient transport from the subarctic gyre to the subtropical gyre through the KOTA. In addition to this horizontal transport, the upward flux of nutrients from North Pacific Intermediate Water to lighter densities that is caused mainly by turbulent mixing may be critically important in the subtropical gyre. Sarmiento et al. (2004) suggested that North Pacific Intermediate Water is expected to have an important role in surface biological activity at lower latitudes. However, as latitudinal and vertical coverage of the measurements are still limited to examining the role of vertical mixing of basin-scale nutrients, the location of major turbulent nitrate transport from North Pacific Intermediate Water in the subtropical gyre has been unclear. Such transport could constitute part of the inter-gyre transport of nitrate.
To reveal the basin-scale distribution of the vertical turbulent nitrate flux in the western North Pacific, we made concurrent observations of turbulence and nitrate along three north-south transects across the subarctic gyre and the subtropical gyre (Fig. 1) from the surface down to about 2000 m for turbulence and down to about 1000 m for nitrate. The rest of the paper is organized as follows: In Sect. 2, details of observations and data processing methods are presented. In Sect. 3, hydrographic cross-sections including nitrate concentration and its gradient, vertical diffusivity, and the vertical turbulent nitrate flux are presented. We also distinguish the contributions of the nitrate gradient and diffusivity to the nitrate flux. Finally, in Sect. 4, we compare the measured levels of turbulence and the nitrate flux with those of previous studies and discuss the influence of the vertical nitrate flux on the ecological biogeography of the western North Pacific.
To distinguish the general hydrographic features of the above observation stations, we mapped the stations onto the absolute dynamic topography field. The absolute dynamic topography data with a horizontal interval of 1/4° were processed by Segment Sol multimissions d'ALTimetrie

Data processing and analysis
The ISUS measurements of nitrate concentration during the downcast of CTD surveys were used in the present study. The raw data of the downcasts obtained with a vertical interval of 1 dbar were linearly adjusted to the nitrate concentration of the water samples taken during the upcasts. The limit of detection of the ISUS is ~ 1.5 mmol m −3 (e.g., Johnson and Coletti 2002). Near the surface, we treated the nitrate concentration as 0 mmol m −3 at depths shallower than those where the concentration obtained from the water samples (around 0, 5, 10, 20, 30, 40, 50, 75, 100, 125, 150, and 200 dbar at each station) became smaller than 0.03 mmol m −3 . Using the data obtained from water sampling, we also confirmed that these adjustments represented the distribution of nitrate, including the small gradient of nitrate around the depth of the base of the euphotic zone (as defined below), especially in the oligotrophic subtropical region. Chlorophyll a fluorescence measured by the fluorometer was also linearly adjusted to the chlorophyll a concentration of the water samples determined by a Turner Designs fluorometer (Welschmeyer 1994). The base of the euphotic zone was defined as the depth where the photosynthetically active radiation was reduced to 1% of that at 5 dbar at Sts. 1-3, 5, 6, 8-12, 16, and 20. The depth of the base of the euphotic zone was also estimated using the method of Lee et al. (2007) for all stations.
We calculated turbulent kinetic energy dissipation rates ε (W kg −1 ) from microscale velocity measurements by the VMP using the same method as Kaneko et al. (2012), which yields ε profiles with an approximate vertical resolution of 10 dbar. Vertical diffusivity K ρ (m 2 s −1 ) was calculated from ε and the squared buoyancy frequency N 2 (s −2 ) as assuming a mixing efficiency Γ = 0.2 (Osborn 1980).
Considering the intermittent nature of turbulence, we estimated the vertical turbulent nitrate flux F N as follows, assuming that eddy diffusivity, K ρ , is equivalent to the vertical diffusivity of nitrate, K N : where C N indicates the nitrate concentration (mmol m −3 ) measured by the ISUS, and overbars denote averages over a 50-m grid-bin in the vertical direction. Here, since large variability would be expected because of the turbulence, we employed 50 m grid-bins based on the method of Kaneko et al. (2013) to improve the accuracy of the estimation. Note that K is calculated as follows: Confidence intervals at the 95% level were also estimated using the bootstrap method. We calculated the 95% confidence levels for both the vertical diffusivity and the gradient of nitrate and compared their influence on the confidence level of the fluxes. The results showed that large variability in the diffusivity had more impact on the confidence level of the fluxes than that of the nitrate gradient (not shown). We then calculated the vertical flux divergence as the difference in flux between adjacent 50 m bins.

Factors driving turbulence
Tides and winds are the major forcing factors of turbulence. They enhance internal waves at tidal and near-inertial frequencies, as well as directly dissipating kinetic energy in the bottom and surface boundary layers, respectively. Thus, they were estimated for comparison with turbulence. As an indicator of the internal-tide forcing in the local region, we employed the 0.5° resolution dataset of temporally and vertically averaged dissipation rate caused by the internal tide provided by de Lavergne et al. (2019), denoted as D tide . As an indicator of the surface wind forcing in the local region, we estimated the local surface energy flux from the wind field as E wind following Oakey and Elliott (1982), that is, where ρ a is the air density (taken as constant; 1.2 kg m −3 ), C 10 is the drag coefficient (1.3 × 10 −3 ), and U 10 is wind speed at 10 m. We used the third-generation Japanese Ocean Flux Data Sets with Use of Remote Sensing Observations (J-OFURO3; Tomita et al. 2018; https ://www.j-ofuro .scc.utokai .ac.jp) for analysis of the daily wind speed data at each VMP station. Note that although the energy from the wind dissipates in the mixing layer near the surface, some of the energy is transferred to the near-inertial waves and can affect dissipation within the interior of the ocean. It is also important to note that E wind does not represent the near-inertial waves generated by the frontal processes reported in previous studies (e.g., Alford et al. 2013;Nagai et al. 2015).

Hydrographic structure and the intensity of turbulence
During the period of observation, the cyclonic subarctic gyre and the anticyclonic subtropical gyre were clearly recognized in the absolute dynamic topography field around the observational stations (Fig. 1). Along the 155° E transect, the northernmost station (St. 12) was located in the subarctic gyre, and the southernmost station (St. 22) was in the North Equatorial Current, which is the southern bound of the subtropical gyre (Table 1). It is noted that the southern stations  in the subtropical gyre are located over the characteristic topography (Fig. 2). Stations 1-5 and 11 along the 160° E transect appear to be distributed from the subarctic gyre to the KOTA. The KE, generally referred to as the northern bound of the subtropical gyre, did not flow simply eastward but took a Z-shaped path around 155° E, and was thus observed in a relatively wide latitudinal range of 31°-36° N, where three stations (T1, 16, and 17) were located (Fig. 1). Here, the subarctic front and the KE were detected based on the temperature distribution (4 °C at 100 m; Favorite et al. 1976, and 14 °C at 200 m;Kawai 1969, respectively;Fig. 3a-c). The four stations 6-9 along the ESM transect were located above the sea mounts with depths of 1150-2390 m (Fig. 2). In addition to the properties at the above horizontal transects, stratifications of different water masses were also captured (Fig. 3). One of the most prominent features was the North Pacific Intermediate Water that is characterized by the vertical salinity minimum centered at 26.8 σ θ (Yasuda 1997). In the northern and middle parts of the subtropical gyre, from the KE (~ 36° N, just south of the axis) to ~ 19° N, a thick layer (i.e., a pycnostad) at 25.0-25.5 σ θ (Fig. 3d) was observed above the North Pacific Intermediate Water, suggesting subtropical mode water (Masuzawa 1969). A shallow pycnocline was observed within the North Pacific Tropical Water (Cannon 1966), characterized by high salinity in layers around 24.0 σ θ ; this occurred in the southern part of the subtropical gyre from ~ 24° N to the North Equatorial Current (~ 11° N) (Fig. 3d).

Nitrate flux and chlorophyll a concentrations
Nitrate distribution resembled that of density, with the well-known contrast between oligotrophic subtropical and non-oligotrophic subarctic waters in the near surface ( Fig. 8a-c). In the KE, the local downward concave nitrate distribution associated with isopycnals occurred at depths deeper than 200 dbar (along 26.0 σ θ ) around 35° N (Fig. 8a). In addition to the large-scale bowl-shaped vertical maximum at mid-depths (26.0-27.0 σ θ along 155° E), the vertical gradients of nitrate were characterized by nitraclines at the subsurface along the 24.0-25.0 σ θ isopycnals from the KE to the center of the subtropical gyre (Fig. 8d). The depth of the subsurface nitracline with a vertical gradient of nitrate exceeding 0.05 mmol m −4 increased with decreasing latitude. The mid-depth nitracline in the subtropical gyre situated at 400-800 dbar was associated with the vertical salinity minimum corresponding to the North Pacific Intermediate Water. The intensity of the nitracline around 26.5 σ θ exceeded 0.1 mmol m −4 in the subarctic region, including the ESM (Fig. 8e, f).
In addition to that found in the subarctic gyre, a high concentration of chlorophyll a (> 0.5 mg m −3 ) was observed in the KOTA (Fig. 9d, e). A similar concentration to that in the KOTA (0.5 mg m −3 ) was also recognized at Sts. 16-17 in the KE (Fig. 9d). Chlorophyll a elevation was observed as a subsurface chlorophyll a maximum, especially south of the   (Fig. 10a-c). The vertically integrated chlorophyll a concentration was also higher in the subarctic gyre than in the subtropical gyre, consistent with that at the subsurface chlorophyll a maximum (Fig. 10a-c). The magnitude of the vertically integrated chlorophyll a concentration in the KOTA was at a similar level to that in the subarctic gyre, suggesting prominent peaks of chlorophyll a concentration at the subsurface chlorophyll a maximum in the region (e.g., St. 3).

Vertical flux at the base of the euphotic zone
There was a large contrast in fluxes at the base of the euphotic zone between the subtropical gyre and other regions (Fig. 10d-f). In the subtropical gyre, the fluxes were estimated as O(10 −8 ) mmol m −2 s −1 or smaller, owing to the weak vertical gradient of nitrate (Fig. 9a). The small gradient of nitrate was also confirmed by the nitrate concentrations obtained from water sampling in the subtropical gyre. In contrast to the small fluxes in the subtropical gyre, in the subarctic gyre, the fluxes were O(10 −7 ) mmol m −2 s −1 . In the ESM, although their magnitude varied considerably, the fluxes were O(10 −6 ) mmol m −2 s −1 . Fluxes in the KOTA Error bars indicate 95% confidence intervals. VIC means vertically integrated chlorphyll a. were at the same level as in the subarctic gyre, O(10 −7 ) mmol m −2 s −1 . The relationship between the fluxes and the vertical gradient of nitrate at the base of the euphotic zone ( Fig. 11a) was more ambiguous (r = 0.36, p = 0.15) than that between the fluxes and diffusivity, where significant correlation (r = 0.84, p < 0.01) was found (Fig. 11b), suggesting a large contribution from the diffusivity to the fluxes. In particular, high diffusivities in the ESM corresponded to large fluxes. Although fluxes at the base of the euphotic zone did not show a significant relationship with the subsurface chlorophyll a maximum concentration when all stations were considered, the relationship became clearer (r = 0.45, p = 0.096, Fig. 12a) when Sts. 3 and 11 were excluded. Note that there is a possibility of missing flux enhancement due to short-term mixing when only a single measurement of turbulence is made, which might lead to the ambiguous relationship when all stations are considered. In addition, at Sts. 3 and 11, a prominent chlorophyll a maximum was observed almost at the surface (Fig. 9e), suggesting that a prominent near-surface bloom occurred associated with a near-surface nitracline. At the repeated station (Sts. 2 and 11), the nearsurface chlorophyll a peak disappeared rapidly and became a subsurface chlorophyll a maximum in the 2 weeks (Fig. 9e).
Although the cause of such a prominent subsurface chlorophyll a maximum remains unclear, such a surface bloom in the KOTA might affect the relationship between the fluxes and subsurface chlorophyll a maximum concentration.
In the case of the vertically integrated chlorophyll a concentration, a moderate relationship was obtained (r = 0.43, p = 0.082, Fig. 12b). The relationship between the fluxes and the vertically integrated chlorophyll a concentration might indicate a longer-term contribution of vertical fluxes to the near-surface production as a result of the accumulation of nutrient supply from the deeper layer to the surface and subsequent production in the euphotic zone, assuming that the level of fluxes remains the same. The result showed mitigation of outliers in the KOTA, compared with those shown in Fig. 12a. This suggests the importance of verifying the time-scale of observations in future work for more accurate estimation, especially in the KOTA. In conclusion, for the present observations, although there was not such a clear linear relationship between the fluxes and chlorophyll a concentration when all stations were considered, the contrast between the subtropical and the subarctic gyres including the ESM were substantial (Fig. 10).

Flux divergence at greater depths
Divergence of the vertical fluxes was clearly observed at mid-depths in the region of the KOTA and the KE along the 155° E section (Sts. 14-16), and at depths from the surface to the subsurface in the ESM (Fig. 13). Along the 155° E section, intense divergence/convergence [O(10 −8 ) mmol m −3 s −1 ] was identified in the upper part of the North Pacific Intermediate Water (26.0-26.8 σ θ , Fig. 13a). Prominent divergence/convergence around the North Pacific Intermediate Water density was also observed at 600 dbar beneath St. 19 (subtropical gyre), and at 400 dbar beneath St. 1 (KOTA), where enhancement of fluxes associated with the elevation of diffusivity was evident. In the ESM, divergence/convergence peaks of O(10 −7 ) mmol m −3 s −1 were observed at depths down to 300 dbar (Fig. 13c). The density at these depths almost reached 27.0 σ θ , suggesting that more direct transport of nitrate from these depths to the surface occurred in the ESM.

Discussion
In this study, we used direct measurements to estimate nitrate fluxes in the western North Pacific. The observed magnitudes vary considerably, and a flux at the base of the euphotic zone of O(10 −6 ) mmol m −2 s −1 was identified in the ESM due to active internal tides (Fig. 10f). In other parts of the subarctic gyre, fluxes of 1-8 × 10 −7 mmol m −2 s −1 were observed at the base of the euphotic zone (Fig. 10d, e). In contrast, the magnitude of the fluxes at the base of the euphotic zone in the subtropical gyre was estimated to be smaller than 1 × 10 −7 mmol m −2 s −1 (Fig. 10d). In the frontal areas, including the KE and the KOTA, the fluxes at the base of the euphotic zone were O(10 −7 to 10 −6 ) mmol m −2 s −1 (Fig. 10d, e). In the eastern North Pacific, Hayward (1987) estimated the turbulent nitrate flux in the subsurface as 0.4-1.6 × 10 −7 mmol m −2 s −1 in the subtropical region by multiplying diffusivity and nitrate distributions obtained from non-simultaneous observations. Fernández-Castro et al. (2015) estimated nitrate diffusive fluxes due to salt fingers plus mechanical turbulence in the eastern part of the North Pacific Tropical Gyre as 1.8 ± 2.2 × 10 −6 mmol m −2 s −1 . The level of fluxes at the base of the euphotic zone observed here in the subtropical gyre is the same as that reported by Hayward (1987) but smaller than that estimated by Fernández-Castro et al. (2015). The difference in the fluxes between those of the present study and those from Fernández-Castro et al. (2015) is caused by both the difference of the vertical gradient of nitrate (0.011 ± 0.037 and 0.142 ± 0.084 mmol m −4 , respectively) and that of the vertical diffusivity (0.14 ± 0.13 × 10 −5 and 1.8 ± 2.8 × 10 −5 m 2 s −1 , respectively). Note that the values for the present study are from Sts. 16 to 22; for vertical diffusivity, we used the center value of the bootstrap method at each station for comparison with the previous study, and the 95% confidence interval of the stations ranged from 0.046 to 0.59 × 10 −5 m 2 s −1 . In contrast to the observations of Fernández-Castro et al. (2015) conducted in March, our observations were made from August to September. Thus, the development of a deeper nitracline and calm weather conditions in summer might produce the weaker gradient of nitrate and lower diffusivity near the surface observed in the present study. The level of turbulent nitrate flux in the subtropical gyre in the present study, O(10 −8 ) mmol m −2 s −1 , is one order smaller than that of nitrate fixation (another dominant source of nitrate in oligotrophic regions) reported by Shiozaki et al. (2009) (2.94-17.6 × 10 −7 mmol m −2 s −1 ) in the subtropical oligotrophic North Pacific along 155° E (16°-28° N), although their observations were made in winter. Fernández-Castro et al. (2015) reported that nitrate diffusion dominates nitrate fixation in the eastern part of the North Pacific using concurrent observations; thus, further studies, especially those in the same season, are required to draw firm conclusions. The magnitude of O(10 −7 to 10 −6 ) mmol m −2 s −1 that was estimated in the frontal areas including the KE and the KOTA is of the same order as that of new production in the KE in August [O(10 −6 ) mmol m −2 s −1 ] estimated from data provided by Yokouchi et al. (2006), assuming the Redfield ratio of C:N = 106:16 (where C is carbon and N is nitrogen). At greater depths (higher densities), relatively large flux divergence/ convergences were observed in the ESM (Fig. 13c). We, therefore, calculated a statistical mean of the divergence in each 0.5 σ θ range in the region (Fig. 14a). Then, divergence in 26.5-27.0 σ θ and convergence in 25.5-26.5 σ θ were estimated. As the isosurface of 25.5 σ θ was located almost at the surface in the ESM (Fig. 3), the convergence in 25.5-26.5 σ θ may contribute to nutrient transport from the density of North Pacific Intermediate Water to the surface. In addition, substantial flux divergence was estimated in the KOTA and KE (Fig. 13a), with flux divergence (convergence) in the density range 26.5-27.0 σ θ (25.5-26.5 σ θ ) (Fig. 14b). Thus, we also calculated a statistical mean of the divergence for Sts. 14-16 (Fig. 14b). Although the mean showed divergence in 26.5-27.0 σ θ that corresponds to the core density of North Pacific Intermediate Water, the mean in 25.5-26.5 σ θ was evaluated as convergence [O(10 −9 ) mmol m −3 s −1 ]. Convergence due to turbulent diffusivity can be regarded as nutrient transport from the core of North Pacific Intermediate Water to lighter density in the subtropical gyre. In addition, if the convergence reaches up to the core density of subtropical mode water, it may play an important role in nutrient transport from the North Pacific Intermediate Water to the lower latitude surface (Sarmiento et al. 2004) because subtropical mode water is connected to the surface through wintertime mixing. A more detailed evaluation of the nutrient transport from the North Pacific Intermediate Water to subtropical mode water is required. The flux observed at the upper boundary of the subtropical mode water, O(10 −7 ) mmol m −2 s −1 (Fig. 9a, St. 18), is far smaller than that indirectly estimated from profiling float data by Sukigara et al. (2011), 9 × 10 −6 mmol m −2 s −1 . Because the vertical gradient of nitrate is similar (~ 0.05 mmol m −4 ) in the present study and that of Sukigara et al. (2011), the difference in diffusivity between them is responsible for the difference in the fluxes. As the measurement of turbulence at the stations was conducted only once, the seasonal-and monthly-scale variabilities of turbulence, including heavy weather conditions, may explain the difference. In addition, shorter timescale variation caused by the internal tides can also affect the estimation. Since there was only a single deployment at each station, we might have missed enhancements of turbulence due to short-term internal tides that occur on timescales of a few hours. Other possible reasons for the difference include the fact that the observations were made in different seasons (from August to September in this study and May to July in Sukigara et al. 2011) and that Sukigara et al. (2011) assumed a one-dimensional balance of oxygen for estimating diffusivity.
To discuss the energy source of the observed dissipation, the vertically averaged dissipation rate throughout the water column, ⟨ ⟩ WC was estimated at each station. We employed E wind with zero lag and the nearest grid data from the deployment after confirmation of the lag-correlation analysis (Supplemental Fig. 1). Latitudinal distributions of ⟨ ⟩ WC , E wind , and D tide along each transect are shown in Fig. 15. In addition to those in the ESM, elevations of D tide (> 10 −3 W m −2 ) were also estimated in the lower latitudes of the subtropical gyre (< 30° N), where characteristic topography existed (Fig. 2). In the subtropical gyre region < 30° N, ⟨ ⟩ WC was also elevated (> 2 × 10 −10 W kg −1 ). In contrast, along 160° E, although D tide was small (< 0.4 × 10 −4 W m −2 ), ⟨ ⟩ WC had relatively high values, implying a wind influence, especially at 40° N and 44° N. At 47° N near the subarctic front, although there was a wider range of uncertainty in the turbulence, the elevation of ⟨ ⟩ WC was suggested, whereas E wind and D tide were small. Thus, near-inertial waves may be responsible. In another frontal region, at St. 15, ⟨ ⟩ WC was elevated (> 4 × 10 −10 W kg −1 ) in spite of weak E wind and D tide , suggesting turbulence enhancement due to the near-inertial waves near the KE as mentioned by Kaneko et al. (2012) and/or Nagai et al. (2015). Weather conditions were calm before the turbulence measurement at St. 15, with neither developed cyclones nor typhoons (e.g., mean wind intensity for 20 days and its standard deviation at 10 m estimated from the J-OFURO3 data was 6.4 ± 2.1 m s −1 ), implying frontal processes as a possible source of the turbulence (Alford et al. 2013;Nagai et al. 2015) in addition to the near-inertial waves propagating from higher latitudes (e.g., Garrett 2001;Alford 2003). These near-inertial waves caused by the frontal processes may lead to the large discrepancy between dissipation and E wind . In total, there was a significant positive correlation between ⟨ ⟩ WC and D tide , and between ⟨ ⟩ WC and E wind (r = 0.43, p < 0.05, and r = 0.55, p < 0.01, respectively, not shown). We divided the data into three groups according to the magnitude of D tide ,; the upper third as the high-D tide group (Sts. 6,8,9,19,20,21,and 22), the next third as the moderate-D tide group (Sts. 7,10,12,16,17,18,and T1), and the lower third as the low-D tide group (Sts. 1,3,5,11,13,14,and 15). Then, an analysis of covariance was conducted with E wind as the covariate and ⟨ ⟩ WC as the dependent variable. Although the slopes of the regression lines were not significantly different [F(2, 15) = 0.58, p = 0.57], there was a significant effect of D tide on levels of ⟨ ⟩ WC after taking account of the effect of E wind [Fig. 16, F(2, 17) = 5.71, p < 0.05]. Therefore, although dissipation in the whole water column would be affected by the surface wind, the difference due to the level of tidal mixing was also significant.
This study has shown how the enhancement of fluxes due to internal tides in the ESM compares with other subarctic regions. The ESM region is expected to contain hotspots of vertical redistribution of heat and materials, including nutrients. In the subtropical gyre, where there is thought to be a weak nitracline and elevated diffusivity due to internal tides, fluxes at the subsurface can be strengthened if the vertical gradient of nitrate is increased locally. Such an increase in the gradient can be temporarily caused by cyclonic eddies such as at St. 17 (Figs. 1b, 3d). Thus, frontal areas in the subtropical gyre, such as the subtropical counter currents where active eddy activity was reported (e.g., Qiu and Chen 2011), might be expected to act as a nutrient transport path in the subtropical gyre, in addition to the KE. The KE and KOTA have been shown to be key regions of nutrient transport from North Pacific Intermediate Water into less dense waters. However, it should be noted that the present study is based on one-time sparse measurements of turbulence that are unlikely to be sufficient to capture the patchy turbulence caused by the short-term internal waves, and the active variability around fronts and eddies in these regions. Therefore, future work will require long-term repeat observations of vertical diffusivity, including mooring measurements and profiling floats equipped with turbulence sensors to clarify Error bars indicate 95% confidence intervals Fig. 16 Scatter plot of vertically averaged dissipation rate against E wind . Error bars indicate 95% confidence intervals. Solid and dashed lines denote the linear regression and confidence interval of the slope with respect to each group, respectively nitrate circulation. As the first step, the basin-scale variability in the vertical nitrate flux provided by the present study will help improve our understanding of nitrate sources for biological productivity in the western North Pacific.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/.