Community composition and photosynthetic physiology of phytoplankton in the western subarctic Pacific near the Kuril Islands with special reference to iron availability

Abstract

The western subarctic Pacific (WSP) is known as one of the most productive regions among the world’s oceans in spring. However, its oceanic waters are also known as a High Nutrient, Low Chlorophyll (HNLC) region during summer due to low iron (Fe) availability in seawater. Indeed, recent studies have demonstrated that the distribution of Fe in the WSP is complex and heterogeneous. This study thus investigated the effects of Fe availability on the community composition and photophysiology of surface phytoplankton from coastal to offshore waters in the WSP in the summer of 2014. Although relatively high concentrations (>2 mg m⁻³) of chlorophyll (chl) a were found in the Sea of Okhotsk and some coastal waters, low chl a concentrations (<1 mg m⁻³) were commonly observed in offshore waters. Based on dissolved Fe and macronutrient concentrations, we deduced that low Fe availability limited phytoplankton growth in offshore waters, whereas low silicate and/or nitrate levels limited growth in the shelf areas. Scanning electron microscopy also revealed that the centric diatom Chaetoceros exclusively dominated the diatom assemblages in the shelf and coexisted with pennate diatoms in offshore waters, respectively. Primary productivity in surface waters was negatively correlated with the bottom of the euphotic layer or the light saturation index of the photosynthesis–irradiance curve, which indicates that the phytoplankton assemblages were well acclimated to in situ light conditions regardless of the water masses.

1. Introduction

Iron (Fe) is a crucial micronutrient for algal photosynthetic processes such as pigment synthesis, photosynthetic electron transport, nitrate reduction, and detoxification of reactive oxygen [Sunda and Huntsman, 1995; Twining and Baines, 2013; Behrenfeld and Milligan, 2013]. Indeed, a number of in situ mesoscale Fe fertilization experiments have revealed that Fe input enhances phytoplankton biomass and primary productivity in High Nutrient, Low Chlorophyll (HNLC) waters [e.g., Boyd et al., 2007].

The western subarctic Pacific (WSP) is the largest CO₂ sink by biological activity among the world’s oceans [Takahashi et al., 2002] and has high vertical transport efficiency of particulate organic carbon in spring [Honda, 2003; Kawakami et al., 2004; 2015]. These distinctive characteristics of the WSP are partly due to the massive spring diatom blooms in the Oyashio region off the Kuril Islands [Isada et al., 2010; Yoshie et al., 2010; Suzuki et al., 2011]. The WSP is also known to be an HNLC region in summer and winter [e.g., Tsuda et al., 2003; Suzuki et al., 2009; Hattori-Saito et al., 2010; Fujiki et al., 2014]. In this region, Fe deficiency for large (> 20 μm) diatoms was also observed in Oyashio waters during the bloom phase [Hattori-Saito et al., 2010] and even in coastal waters between the Kuril Islands in summer [Yoshimura et al., 2010; Sugie et al., 2013; Suzuki et al., 2014].

Regarding Fe sources, Nishioka et al. [2007; 2013; 2014] and Nishioka and Obata [2017] suggested that the North Pacific Intermediate Water (NPIW) supplies terrestrial Fe from the Amur River via the Sea of Okhotsk and the Kuril straits to the Pacific. Nishioka et al. [2011] also pointed out that the Fe supply by upward flux was much larger than atmospheric Fe deposition. Additionally, the temporal variation of Fe concentration in seawater reflected the vertical flux of Fe rather than the dust Fe supply. In addition, the East Kamchatka current derived from the Aleutian Islands and the Bering Sea also flows into the WSP [Favorite, 1976; Stabeno and Reed, 1992; Kono and Kawasaki, 1997; Yasuda, 2003], which can be another Fe source [Suzuki et al., 2002].

The distribution and dynamics of Fe in the WSP are thus highly heterogeneous and complex [Nishioka et al., 2014], and it is thus crucial to unravel Fe-related biogeochemical processes such as phytoplankton photosynthesis in the WSP. At present, our knowledge of Fe availability to phytoplankton in coastal and shelf waters (<2500 m in depth) is quite limited despite this region having unique physical and biogeochemical properties [e.g., Yagi and Yasuda, 2012; Nishioka et al., 2013]. There are two challenges to address the Fe availability to phytoplankton in the WSP near the Kuril Islands: a) little attention has been paid to the taxonomy and physiological responses of phytoplankton to Fe availability to date; and b) there is little available information for the upstream of the Oyashio around the eastern Kamchatka Peninsula regarding biological, physical, and chemical properties.

Therefore, this study examined the effects of Fe availability on the community composition and photophysiology of phytoplankton in the WSP near the Kuril Islands and the eastern Kamchatka Peninsula. Phytoplankton community composition was investigated with scanning electron microscopy (SEM) [e.g., Sugie and Suzuki, 2017], which enabled the identification and quantification of armored cells such as diatoms that were predominant in the study area [Mochizuki et al., 2002; Yoshie et al., 2010; Suzuki et al., 2011, 2014]. Moreover, the photophysiology of phytoplankton was estimated by photosynthesis-irradiance (P-E) curves, which provide powerful insights into the physiological states of phytoplankton such as light-dark acclimation and responses to nutrient availability [Sakshaug et al., 1997; MacIntyre et al., 2002; Bouman et al., 2018]. Spatial patterns of the photophysiology were also investigated with physicochemical parameters (i.e., temperature, macronutrient and Fe concentrations, and light availability).

2. Materials and Methods

2.1. Water sampling and optical observations

Seawater sampling was conducted in the western subarctic Pacific along the Kuril Islands and off the eastern Kamchatka Peninsula aboard the R/V Professor Multanovskiy from 2 June to 8 July 2014 (Fig. 1). Sampling stations spanned from the shallow shelf areas (e.g., bottom depth 88 m at station B6) to the deep offshore regions near the Kuril-Kamchatka Trench (e.g., bottom depth 7,994 m at station A5). Bottom depths at the sampling stations were obtained from 15 arc-second gridded bathymetry data provided by General Bathymetric Chart of the Oceans (GEBCO; https://www.gebco.net/data_and_products/gridded_bathymetry_data/). Station D9 was located in the Sea of Okhotsk, whereas the other stations were in the Pacific Ocean (Fig. 1). Surface seawater samples were obtained from an approximately 5-m layer around noon using a CTD carousel multisampler system (CTD-CMS) with acid-cleaned Niskin bottles washed with trace-metal-clean techniques [Obata et al., 1993; Takeda and Obata, 1995; Nishioka et al., 2001]. At each station, seawater was collected into an acid-cleaned, trace-metal-free 9 L polycarbonate (PC) bottle. The seawater obtained was dispensed into four acid-cleaned 300 mL PC bottles and two polystyrene tubes as primary productivity and macronutrient samples, respectively. Concentrations of macronutrients (nitrate + nitrite, ammonium, phosphate, and silicate) were determined with an autoanalyzer (QuAAtro, BL-Tech Inc.). Mixed layer depths (MLDs) were calculated as the depth at which the potential density anomaly (Δσ_θ) of the water column increased by 0.125 kg m⁻³ relative to the sea surface at 10 m [Monterey and Levitus, 1997]. Optical observations were conducted to acquire in situ vertical profiles of photosynthetically available radiation (PAR, 400–700 nm), E_d(PAR), and the spectra of downward irradiance E_d(λ) using a Compact-Optical Profiling System (C-OPS, Biospherical Instruments Inc.) [Hooker et al., 2013]. The diffuse attenuation coefficient of PAR (K_d(PAR)) was calculated using the NASA Processing of Radiometric Observations of Seawater using Information Technologies (PROSIT) software [Hooker et al., 2018] with the aid of a solar reference for measuring the incident PAR (E₀(PAR)) above the sea surface. Additionally, a LI-190SB air quantum sensor and a LI-1400 data logger (LI-COR, Inc.) were used for the continuous measurement of sea surface PAR irradiance, E₀(PAR). Based on the results for K_d(PAR), euphotic zone depth (Z_eu) was calculated as a 1% light depth compared with the surface PAR [Kirk, 2010]. Mean surface PAR in the daytime was calculated daily by averaging the daytime data from the air quantum sensor, and depth averaged PAR within MLDs was estimated from the vertical PAR profiles based on the K_d(PAR) values.

Figure 1.

Seawater sampling stations during the Mu14 expedition along the Kuril Islands and off the eastern Kamchatka Peninsula.

The black, green, red, and blue closed circles in the map denote Okhotsk, southern shelf (SS), northern shelf (NS), and offshore waters, respectively.

The Okhotsk indicates the only station in the Sea of Okhotsk. Waters on the shelf and shelf slope (<2500 m) were defined as shelf waters. Shelf waters near the southern and northern Kuril Islands were defined as SS and NS. Offshore waters outside of the shelf slope (>2500 m) were defined as offshore waters.

2.2. Size-fractioned chlorophyll a concentration

Seawater was sequentially filtered onto a 47 mm nylon mesh (20 μm pore size), a 47 mm Nuclepore membrane (2 μm pore size), and a Whatman GF/F filter (nominal 0.7 μm pore size) under gentle vacuum (< 0.013 MPa). These filters were immediately flash frozen in liquid nitrogen and stored in a deep freezer (−80°C) until analysis. Pigments were extracted by soaking in N,N-dimethylformamide (DMF) at −20°C for at least 24 h [Suzuki and Ishimaru, 1990]. Concentrations of chlorophyll (chl) a were determined with a Turner Designs 10-AU fluorometer using the non-acidification method of Welschmeyer [1994]. Three size fractions were obtained as microphytoplankton (>20 μm), nanophytoplankton (2–20 μm), and picophytoplankton (0.7–2 μm).

2.3. Microscopy

Seawater was fixed with a buffered neutral formaldehyde solution (final concentration 2%). Fixed samples were stored at 4°C until analysis. A subsample of the fixed sample was filtered onto a 25 mm 0.4 μm Nuclepore membrane filter under gentle vacuum (< 0.013 MPa). The filter was rinsed with deionized water to desalt and then naturally dried. The dried filter was secured with conductive carbon tape on a microscope stage, and the surface of the filter on the stage was coated with Pt-Pb alloy by vapor deposition using an MSP-1S magnetron spatter device (Vacuum device Inc.). Armored protistan cells were counted with a scanning electron microscope (VE-8800, Keyence Co., Ltd.) with a magnification of >1000× according to Nosaka et al. [2014; 2017]. Species identification was based on Tomas [1997], Round et al. [2007], Konno et al. [2007], and Medvedeva and Nikulina [2014]. For identification of Chaetoceros species, the number of plastids in a cell was enumerated under bright field microscopy (Keyence BZ-9000). Diatom size (i.e., apical length or valve diameter of diatoms) was measured if >20 cells of a given diatom could be observed using the software ImageJ (National Institutes of Health). The abundance of armored protists in this study was expressed as the number of cells per the volume of seawater.

2.4. Variable chlorophyll a fluorescence

To assess the photochemical efficiency of photosystem II (PSII), variable chl a fluorescence was measured with Pulse Amplitude Modulation (PAM) fluorometry. Seawater was dispensed into a 30 mL shading polyethylene bottle and stored in an incubator at in situ temperature for 30 minutes for dark acclimation to open the reaction centers of algal PSII. After the incubation, the maximum quantum yield of PSII photochemistry (F_v/F_m) for phytoplankton was measured with a Water-PAM (Walz) following Liu et al. [2009]. Variable chl a fluorescence measurements were conducted ≥3 times for each sample.

2.5. Light absorption coefficient of phytoplankton

Seawater was filtered onto a GF/F filter (Whatman) with gentle vacuum (<0.013 MPa). Duplicate filters were carefully wrapped with aluminum foil to avoid creases and stored in a deep freezer (−80°C). After the filters were thawed, light absorption measurements were conducted with a multipurpose spectrophotometer (MPS-2450, Shimadzu) equipped with an end-on type photomultiplier tube. Optical densities of particles and detritus on a filter (OD_fp(λ) and OD_fd(λ), respectively) were measured three times per sample at wavelengths between 350 and 800 nm in a 1 nm pitch with a methanol method [Kishino et al., 1985]. All spectra were normalized by subtracting the average of optical densities between wavelengths 730 nm and 750 nm from each sample following Babin and Stramski [2002]. Optical densities were converted to absorption coefficients a_p(λ) and a_d(λ) for particles and detritus (non-algal particles), respectively, with correction for the pathlength amplification factor proposed by Cleveland and Weidemann [1993]. Then, a_ph(λ) calculated as a difference between a_p(λ) and a_d(λ) was normalized to chl a concentration and spectrally weighted for PAR of the incubator lamp for the P-E curve experiment to obtain a mean chl a-specific absorption coefficient of phytoplankton (ā*_ph) in accordance following Isada et al. [2013].

2.6. Photosynthesis vs. Irradiance (P-E) curve experiment and photosynthetic parameters

Seawater was dispensed into twelve 275 mL polystyrene bottles, and a solution of NaH¹³CO₃ (99 atom% ¹³C purity, Cambridge Isotope Laboratories, Inc.) was added to the bottles, which was the equivalent level of ~10% of total inorganic carbon, except for 2 bottles that were used as time-zero samples. A single set of the ¹³C-labeled seawater in the bottles was incubated for 2 h in a temperature-controlled incubator under 10 different light intensities from 1.44 to 3000 μmol photons m⁻² s⁻¹ at in situ temperature. After the incubation, seawater samples were filtered onto precombusted 25 mm GF/F filters (Whatman) with gentle vacuum (< 0.013 MPa) and stored in a deep freezer (−80°C) until analysis. Following Hama et al. [1983], photosynthetic rates were calculated from uptake rates of ¹³C by phytoplankton measured with an on-line element analyzer (FlashEA1112, Thermo Finnigan)/Isotope ratio mass spectrometer (EA/IRMS) (Delta-V, Thermo Finnigan). After the normalization with chl a concentration, a P-E curve was fitted with the model of Platt et al. [1980]:

$$\begin{gathered} P^{\text{B}}=P_{\text{s}}^{\text{B}}\left[1-\exp \left(-{\alpha }^{\text{B}}\cdot E_{\text{PAR}\text{inc}}/P_{\text{s}}^{\text{B}}\right)\right]\cdot \exp \left(-{\beta }^{\text{B}}\cdot E_{\text{PAR}\text{inc}}/P_{\text{s}}^{\text{B}}\right),\text{ and } \\ {P}_{\max }^{\text{B}}=P_s^{\text{B}}\cdot \left[{\alpha }^{\text{B}}/\left({\alpha }^{\text{B}}+{\beta }^{\text{B}}\right)\right]\cdot {\left[{\beta }^{\text{B}}/\left({\alpha }^{\text{B}}+{\beta }^{\text{B}}\right)\right]}^{\left({\beta }^{\text{B}}/{\alpha }^{\text{B}}\right)} \end{gathered}$$ (1)

The descriptions of P-E curve parameters are presented in Table 1. The maximum quantum yield for carbon fixation (Φ_Cmax) was also calculated:

$${\text{Φ}}_{\text{Cmax}}=0.0231{\alpha }^{\text{B}}/{{\overline{a}}^{\ast }}_{\text{ph}}$$ (2)

Table 1.

Summary of terms and abbreviations of photosynthetic parameters used for P-E curves

Symbol	Definition	Unit
αB	Chl a-normalized initial slope of P-E curve	(mg C mg Chl a−1 h−1) (μmol photons m−2 s−1)−1
βB	Chl a-normalized photoinhibition index of P-E curve	(mg C mg Chl a−1 h−1) (μmol photons m−2 s−1)−1
PBmax	Chl a-normalized maximum photosynthetic rate	mg C mg Chl a−1 h−1
Ek	Light saturation index	μmol photons m−2 s−1

All procedures for the measurements were according to Isada et al. [2013].

2.7. Primary productivity

Seawater in four acid-cleaned 300 mL PC bottles was incubated under ambient temperature in an on-deck temperature-controlled incubator. Prior to incubation, the NaH¹³CO₃ solution was inoculated into all bottles in the same way as in the P-E curve experiment. Three out of the four bottles were incubated for 24 h under natural sunlight, whereas the other was used as a dark bottle. After the incubation, ¹³C uptake rates were determined following the P-E curve experiments (Section 2.7). Primary productivity, PP [mg C m⁻³ d⁻¹], was calculated following Isada et al. [2013].

2.8. Fe analysis

The dissolved Fe (DFe) data in this study are cited from Nishioka et al. [submitted]. Briefly, DFe samples were collected from each station and prefiltered with a 0.2 μm Acropak filter (Pall Co.), followed by the addition of ultrapure HCl (Tamapure AA-10, Tamapure Co., Ltd.) to solubilize all labile Fe in seawater following Nishioka et al. [2001]. Samples were buffered with 10 M formic acid and 2.4 M ammonium formate buffer to adjust the pH to 3.2 just before measurements at an onshore laboratory. DFe concentrations were determined with an Fe(III) flow injection analytical system (Kimoto Electric, Ltd.) after preconcentration using a chelating resin following the chemiluminescence method [Obata et al., 1993; 1997]. The DFe measurements were quality controlled using SAFe and GEOTRACES reference standard seawaters. The consensus values for Fe(III) in the reference standard seawaters are 0.093 ± 0.008 nM (SAFe S) and 1.00 ± 0.10 nM (GEOTRACES GD) (www.geotraces.org), and we obtained values of 0.092 ± 0.008 nM (n = 11) (SAFe S) and 0.939 ± 0.062 nM (n = 9) (D2) using our method. This good agreement demonstrates that our data were comparable with the global GEOTRACES dataset. The detection limit (three times the standard deviation of the Fe(III) concentration of purified seawater that had been passed through an 8-quinolinol resin column three times to remove Fe) was 0.025 nM.

2.9. Statistical analyses

Spearman’s correlation analysis was performed using SigmaPlot ver. 11.0 (SysStat Software, Inc.) to investigate relationships between variables by assuming heteroscedasticity. Cluster analysis was performed to clarify differences in the community composition of armored protists at the genus level among stations. Additionally, redundancy analysis (RDA) was performed to evaluate relationships between physicochemical properties and community composition of diatoms at the genus level following Endo et al. [2018] (see also section 2.10). The cluster and RDA analyses were performed in the statistical software R (http://www.r-project.org). A dendrogram was generated for the cluster analysis using the data of community composition of phytoplankton at the genus level using the Bray-Curtis dissimilarity and Ward’s clustering method. The validity of the RDA analysis was confirmed with the lengths of the gradients of the physicochemical parameters being sufficiently small (<3 standard deviations) assessed using the detrended correspondence analysis (DCA). The Hellinger transformation [Legendre and Gallagher, 2001] was performed for the normalization of diatom community composition data. Ordinations of physicochemical parameters and diatom genera were plotted onto the RDA coordinate plane.

3. Results

3.1. Hydrography

Sea surface temperature and salinity (SST and SSS, respectively) were generally low and ranged between 2.67–9.20°C and 32.3–32.9, respectively (Table 2). Relatively low temperatures were observed near the Bussol’ Strait (2.67°C at station A2 and 3.60°C at station OP3) (Fig. 1 and Table 2). Nitrate and phosphate concentrations were relatively low near the Kamchatka Peninsula (0.31 μM NO₃ and 0.19 μM PO₄ at station PS1; 1.68 μM NO₃ and 0.29 μM PO₄ at station PS5) (Fig. 1 and Table 2). Silicate concentrations at stations near the Kamchatka Peninsula (PS1 and PS5) and the northern Kuril Islands (B5 and B6) were one order of magnitude lower than those at the other stations. Mixed layer depths ranged from 10.0 to 28.5 m, whereas Z_eu ranged from 26.7 to 47.6 m. At all sampling stations, MLDs were shallower than Z_eu. (Table 2).

Table 2.

Sampling conditions and hydrographic and optical data in surface water at 5 m during the Mu14 expedition.

Station	Date	Water depth	SST	SSS	NO3	NO2	NH4	PO4	SiO4	DFe	MLD	Zeu	Fv/Fm
		(m)	(°C)		(μM)	(μM)	(μM)	(μM)	(μM)	(nM)	(m)	(m)
D9 (Okhotsk)	2014.06.08	3311	6.33	32.36	2.03	0.11	0.13*	0.35	20.78	0.71	24.1	29.8	0.370
A2 (SS)	2014.06.09	2099	2.67	32.67	9.57	0.17	0.29	0.96	28.32	0.13	19.5	26.7	0.509
OP3 (SS)	2014.06.29	1065	3.60	32.39	12.14	0.18	0.85	1.35	28.32	0.14	18.2	46.2	0.153
PS1 (NS)	2014.06.14	2118	5.60	32.40	0.31	0.06	UD	0.19	0.61	0.21	17.2	29.5	0.353
PS5 (NS)	2014.06.15	2245	5.50	32.28	1.68	0.08	0.25	0.29	2.32	0.21	10.0	29.8	0.444
B5 (NS)	2014.06.21	1477	7.20	32.68	3.75	0.11	0.33	0.69	3.53	0.07	14.0	45.5	0.270
B6 (NS)	2014.06.22	88	6.80	32.36	4.89	0.11	0.71	0.79	3.19	0.69	17.2	45.4	0.440
C3 (Offshore)	2014.06.16	4934	6.40	32.87	20.70	0.25	0.50	1.83	34.43	0.22	20.9	44.6	0.176
C5 (Offshore)	2014.06.18	3325	6.05	32.77	9.93	0.16	0.39	1.09	12.91	0.05	28.5	40.3	0.304
B3 (Offshore)	2014.06.20	5778	6.80	32.82	18.87	0.24	0.66	1.76	30.9	0.04	23.8	47.6	0.232
A5 (Offshore)	2014.06.28	7994	6.80	32.81	16.64	0.27	0.30	1.53	14.88	0.06	12.8	36.2	0.256
A6 (Offshore)	2014.06.30	5792	9.20	32.77	17.45	0.24	0.35	1.66	18.50	0.03	14.7	41.4	0.314

The Okhotsk indicates the only station in the Sea of Okhotsk. Waters on the shelf and shelf slope (<2500 m) were defined as shelf waters. Shelf waters near the southern and northern Kuril Islands were defined as SS and NS. Offshore waters outside of the shelf slope (>2500 m) were defined as offshore waters. Water depths were obtained from General Bathymetry Chat of the Ocean (see section 2.1 for details). SST: Sea surface temperature; SSS: Sea surface salinity; NO₃, NO₂, NH₄, PO₄, and SiO₂: Concentration of nitrate, nitrite, ammonium, phosphate, and silicate, respectively; DFe: Dissolved iron concentration (cited from Nishioka et al., submitted); MLD: Mixed layer depth; Z_eu: Euphotic layer depth; F_v/F_m: Maximum quantum yield of photosystem II photochemistry determined with PAM fluorometry. Bold indicates that the concentration is lower than the half saturation constants of macronutrients reported in Sarthou et al. [2005] and 0.1 nM DFe.

^*indicates a value below the qualification limit. Qualification limits of nutrients were 0.10 μM NO₃, 0.03 μM NO₂, 0.18 μM NH₄, 0.11 μM PO₄, and 0.95 μM SiO₄.

The sampling stations were classified into 4 domains based on the geographical locations: (i) Okhotsk, (ii) Southern Shelf (SS), (ii) Northern Shelf (NS), and (iv) Offshore (Fig. 1 and Table 2). Station D9 was located in the Sea of Okhotsk. The two shelf waters were defined by the bottom depths, which were shallower than 2500 m, including the shelf slope areas. The shelf water stations A2 and OP3 near the southern Kuril Islands were defined as SS, whereas the shelf stations PS1, PS5, B5 and B6 near the northern Kuril Islands and the Kamchatka Peninsula were regarded as NS. The stations C3, C5, B3, A5 and A6 outside the shelf slope with the bottom depths >2500 m were named Offshore stations (Fig. 1).

Atmospheric conditions during the cruise were generally overcast or foggy except at station PS1 under clear sky. Sea surface PAR irradiance at the sampling time, E₀(PAR), ranged between 632−1755 μmol photons m⁻² s⁻¹. During the expedition, mean PAR in the daytime ranged between 231−661 μmol photons m⁻² s⁻¹. Average PAR in the surface mixed layer ranged between 35.5−101 μmol photons m⁻² s ⁻¹ (Table S2). Although the lowest E₀(PAR) and mean daytime PAR were observed in the Offshore stations (stations B3 and C3, respectively), the lowest PAR in the surface mixed layer was observed at the shallowest depth station B6 in the NS water (Table S2).

3.2. Dissolved Fe (DFe) distribution

Relatively high values (>0.1 nM) of DFe were observed at the NS and SS stations except for station B5, whereas low values were often found in the Offshore stations except for station C3 (Table 2). The highest DFe value (0.71 nM) was observed at station D9 in the Sea of Okhotsk. DFe:NO₃ ratios were generally low at the Offshore (≤ 0.010 nM/μM) and SS (0.012−0.013 nM/μM) stations (Fig. 2). The Okhotsk and NS stations, on the other hand, showed one order of magnitude higher DFe:NO₃ values. At station B5 in the NS, DFe:NO₃ ratio water was exceptionally low (0.020 nM/μM) but was higher than those in the SS and Offshore waters (Fig. 2 and Table S1).

Figure 2.

Spatial variations in the ratios of dissolved Fe to nitrate (DFe/NO₃; nM/μM). Colors indicate values as shown in the color bar (right).

Ratios below 0.043 nM/μM, which correspond to purple color, indicate deficits of DFe relative to nitrate the limiting nutrient, whereas ratios above 0.043 nM/μM, which correspond to blue to red colors, indicate that deficits of nitrate relative to DFe [Nishioka et al., 2011].

3.3. Size-fractioned chl a concentration

Total chl a concentrations, the sum of size fraction chl a data, were generally ≤1 mg m⁻³. Higher chl a concentrations (> 2 mg m⁻³) were observed at station D9 in the Okhotsk, station A2 in the SS, and station B5 in the NS waters (Fig. 3A). Nanophytoplankton was predominant in the Okhotsk (D9) and SS (A2 and OP3) waters (53−72%). Microphytoplankton (>20 μm) was predominant in the NS waters (61−68%) except at station B6 where nanophytoplankton (2–20 μm) dominated (43%) the phytoplankton assemblages. The Offshore waters were mainly dominated by picophytoplankton (0.7–2 μm) (Fig. 3B).

Figure 3.

(A) Concentrations of chl a determined with non-acidification fluorometric analysis based on Welshmeyer (1994) at in situ sampling stations.

Closed black, green, red and blue bars in panel A denote Okhotsk (Ok), southern shelf (SS), northern shelf (NS), and offshore waters, respectively.

(B) Relative contributions of phytoplankton groups to the total chl a at the level of size class determined by size-fractioned chl a measurements with the non-acidification fluorometric method of Welshmeyer (1994) at in situ sampling stations.

Closed, open, and shaded bars indicate microphytoplankton (>20 μm), nanophytoplankton (2−20 μm), and picophytoplakton (0.7−2 μm), respectively.

3.4. Microscopy

Under the SEM, 52 centric diatoms, 17 pennate diatoms, 4 coccolithophores, 10 parmales, and 13 flagellates, including dinoflagellates, were identified (Tables 3 and S4). The centric diatom Chaetoceros species were dominant, particularly C. debilis, C. diadema, and C. furcellatus. (Table 3). Generally, Chaetoceros species were abundant in the Okhotsk, NS and SS (e.g., D9, A2, PS5, B5, and B6 in Fig. 1 and Table 3). The Offshore waters (e.g., stations C3 and A6) were mainly dominated by the pennate diatoms Fragilariopsis and Pseudo-nitzschia species; additionally, the small centric Thalassiosira oceanica was also predominant at station B3, although Chaetoceros somewhat contributed to the offshore community (Tables 3 and S4). Other pennate diatoms Neodenticula seminae and Nitzschia spp. were observed at almost all stations (Tables 3 and S4). At station C5, parmales were highly abundant and diverse, although they were rarely observed at the other stations. Coccolithophores were only abundant at station C5 (Table 3). It was difficult to conduct species identification for the sample obtained at station PS1; however, the sample was filled with intertwined and coagulated setae of Corethron pennatum and Chaetoceros species, including C. atlanticus. Resting spores of Chaeroceros spp. were also detected at the sampling stations except for stations C5 and A6. In particular, 83.7% of Chaetoceros were present as resting spores at station OP3 (Table 3 and S4). Dinoflagellates were abundant at station A5 (17.7%) (Table 3). The cluster analysis showed that NS waters were distinct in terms of the community composition of armored protists from the other waters (Fig. 4). However, phytoplankton assemblages in the other waters were not distinctive among stations. The RDA showed that the first component was explained by the predominance of Chaetoceros and Fragilariopsis (Fig. 5). The contribution of Fragilariopsis had almost an opposite vector to that of DFe. The vector of Cylindrotheca was in almost the same direction as that of NH₄, whereas Neodenticula seminae showed a similar direction to the DFe vector (Fig. 5).

Table 3.

Contributions of armored plankton to the total plankton collected at each sampling station enumerated under scanning electron microscopy.

		D9	A2	OP3	PS5	B5	B6	C3	C5	B3	A5	A6
		(Ok)	(SS)	(SS)	(NS)	(NS)	(NS)	(Off)	(Off)	(Off)	(Off)	(Off)
Coccolithophores						0.06%			3.67%
Parmales		0.38%	0.17%	1.36%		0.39%	2.24%	0.31%	44.4%	0.11%	0.29%
Silicoflagellates		0.08%	0.09%	17.7%				0.50%	0.55%		0.19%
Dinoflagellates				0.45%	0.42%		0.21%	0.19%		0.46%	2.98%	2.25%
Flagellates				1.36%	0.10%	0.58%	3.31%	0.57%		1.77%	0.77%	1.12%
Diatoms		99.5%	99.7%	79.1%	99.5%	99.0%	94.2%	98.4%	51.4%	97.7%	95.8%	96.6%
Diatoms	Size (μm)
Centric diatoms
Asteromphalus spp.	25−180v†				+			+			+
Chaetoceros spp.		++++	++++	+++	++++	++++	++++	+++	++++	+	++++	+
- C. concavicornis	29.3±8.3a	*				*	*		*	*	*	****
- C. convolutus	31.1±4.8a	*			*						*
- C. debilis	11.0±1.7a		*		***	*		**	***		**	***
- C. diadema	23.7±3.2a	*	*	*	**	*	*	**	***	*	***
- C. furcellatus	10.3±1.8a		*		***	*		*	**		*
- Resting spores	-	*	*	****	**	*	*	***			*
- Other Chaetoceros	-	****	****	**	*	****	*****	**	*	****	***
Corethron pennatum	135.7±44.9a	+				+	+	+	+	+	+	+
Minidiscus spp.	> 5v†	+		+				+	+	+	+
Thalassiosira spp.	4.7±1.3v†	+	+	+	+	+	+	++	+	+	++	+++
Pennate diatoms
Cylindrotheca closterium	~17a††				+	+	+	+		+	+	+
Fragilariopsis spp.	12.5±7.8a	++	+	++++	+	+	+	+++	++	++++	++	++
Neodenticula seminae	21.4±3.7a	+	+	+	+	+	+		+	+	+	++
Nitzschia spp.	34.6±9.4a	+	+	+	+	+	+		+	+	+	++
Pseudo-nitzschia spp.	9.8±2.5a	+	+		+	+		+++	+	+	+
Thalassionema spp.	10−200a†			+	+	+		+	+		+	+
Other diatoms	-	+	++	+		+	+	+	+	+	+	++

The contributions reported here were calculated based on the number of cells. Unidentified materials were excluded from the analysis. Ok: Okhotsk, SC: southern shelf, NS: northern shelf, and off: offshore. Bold numbers in the upper six rows indicate planktonic protest group which contributed to >10% of the total cell number at each sampling station. Contributions of diatoms were indicated with the number of plus (+) symbols: +: >0−10%, ++: >10−25%, +++: >25%−50%, ++++: >50%. The contribution of each Chaetoceros species was indicated with the number of asterisks (*) with the same criteria as those of diatoms described above. The contributions of diatoms were independently calculated within diatoms (i.e., excluding other planktonic species; coccolithophores, parmales, silicoflagellates, dinoflagellates, and flagellates). For more details, see Table S4. Size of major diatoms was shown in the second column. The superscripts a and v indicate the apical length and valve diameter to measure the size of a given species, respectively. The † indicates that information of the size was obtained from Tomas (1997) because the cells, often weakly-silicified, were not well preserved or broken, whereas the †† indicates the approximate apical length of the diatom Cylindrotheca closterium that it was difficult to precisely measure due to their weak silicification and twisted body on the filter for SEM.

Figure 4.

(A) Cluster dendrogram of armored plankton community composition (as determined with SEM) among stations. The dendrogram was generated using the Bray-Curtis dissimilarity (shown in the vertical axis) and Ward’s clustering method.

Figure 5.

Coordinate plane of the redundancy analysis (RDA) on the diatom community composition. The validity of the RDA analysis was previously checked with the detrended corresponding analysis (DCA) that the lengths of gradients of environmental parameters were sufficiently small for the RDA analysis (<3SD) The proportions shown in percent indicate contributions of each RDA axis to the total variance (i.e., proportions explaining the total variance). n=12.

3.5. Variable chlorophyll a fluorescence by PAM fluorometry

The average F_v/F_m value at the surface was 0.37 ± 0.11 throughout the expedition (Table 2). Values of F_v/F_m values at stations A2 and OP3 in the SS were the highest and the lowest (0.509 and 0.15) among the sampling stations, respectively (Table 2). There was no significant difference among the NS, SS and Offshore stations (one-way ANOVA, p>0.05; the Okhotsk water was excluded from the statistical test).

3.6. P-E curves and photosynthetic parameters

Values of the initial slope α^B were generally low (~0.02 [(mg C mg Chl a⁻¹ h⁻¹) (μmol photons m⁻² s⁻¹)⁻¹]) except at stations PS1 and PS5 in the NS and OP3 in the SS (Fig. 6A). Values of β^B at stations PS1 and PS5 in the NS were relatively high, and the value at station A6 in the Offshore was conspicuously high (Fig. 6B). The maximum photosynthetic rate P^B_max values were relatively high at stations PS1 and PS5 (13.3 and 8.2 mg C mg Chl a⁻¹ h⁻¹, respectively) in the NS water compared with other stations. The average value of the other stations was 2.18 mg C mg Chl a⁻¹ h⁻¹ and ranged from 1.20 to 3.70 mg C mg Chl a⁻¹ h⁻¹ (Fig. 6C). The average value of light saturation indices, E_k, was 112 μmol photons m⁻² s⁻¹, with the highest at station A6 in the Offshore and the distinctly low value at station OP3 in the SS (Fig. 6D). Values of ā*_ph were highly variable but relatively high values were observed offshore at B3 and A6 in the Offshore stations as well as station OP3 in the SS (Fig. 6E). Maximum quantum yields of carbon fixation, Φ_Cmax, showed high values near the Kamchatka Peninsula (Fig. 6F). The α^B and β^B were positively correlated with F_v/F_m (p<0.01, Table 4). A significant negative correlation was observed between α^B and MLD (p<0.05, Table 5). Additionally, a positive correlation was observed between P^B_max and SST (p<0.01, Table 5). E_k had significant positive relationships with NH₄, Z_eu and daily average PAR irradiances (p<0.05, 0.01, and 0.01, respectively, Tables 5 and S3). Nutrients were all negatively correlated with Φ_Cmax (p<0.05, Table 5), and Φ_Cmax was negatively correlated with SST and SSS as well as with Z_eu (p<0.05, Table 5).

Figure 6.

Photosynthetic parameters at in situ sampling stations.

(A) α^B: Initial slope of a P-E curve [mg C mg Chl a⁻¹ h⁻¹ (μmol photons m⁻² s⁻¹)⁻¹]; (B) β^B: Photoinhibitionindex [mg C mg Chl a⁻¹ h⁻¹ (μmol photons m⁻² s⁻¹)⁻¹]; (C) P^B_max: Maximum photosynthetic rate [mgC mgChl a⁻¹ h⁻¹]; (D) E_k: Light saturation index [μmol photons m⁻² s⁻¹]; (E) ā*_ph: Chl a-normalized light absorption coefficient of phytoplankton [m² mg Chl a⁻¹]; (F) Φ_Cmax: Maximum quantum yield for carbon fixation [mol C mol photons⁻¹]; (G) PP: Primary productivity [mg C m⁻³ day⁻¹]. The black, green, red, and, blue closed bars in each figure denote Okhotsk, southern shelf, northern shelf, and offshore stations, respectively.

Table 4.

Spearman-rank correlation coefficients (ρ) between photosynthetic and environmental parameters data.

	Fv/Fm	αB	βB	PBmax	Ek	ā*ph	ΦCmax	PP
Chl a	−0.37	−0.30	−0.23	0.07	−0.80***	0.71	0.50	0.82***
Fv/Fm		0.73**	0.78**	−0.20	0.36	0.12	0.39	−0.14
αB			0.80***	0.36	0.37	0.00	0.18	0.04
βB				0.41	0.42	0.09	0.17	−0.03
PBmax					0.23	−0.15	−0.37	0.09
Ek						−0.55	−0.67*	−0.73*
ā*ph							0.62*	0.59*
ΦCmax								0.66*

Bold numbers indicate significant relationship between given parameters. Significance levels are denoted by numbers of asterisks:

^*= 0.05,

^**= 0.01,

^***= 0.001, n = 12.

Chl a: Chl a concentration [mg m⁻³]; F_v/F_m: Maximum quantum yield of PSII measured with PAM fluorometry: α^B: Initial slope of a P-E curve [mg C mg Chl a⁻¹ h⁻¹ (μmol photons m⁻² s⁻¹)⁻¹]; β^B: Photoinhibition index [mg C mg Chl a⁻¹ h⁻¹ (μmol photons m⁻² s⁻¹)⁻¹]; P^B_max: Maximum photosynthetic rate [mgC mgChl a⁻¹ h⁻¹]; E_k: Light saturation index [μmol photons m⁻² s⁻¹]; ā*_ph: Chl a-normalized light absorption coefficient of phytoplankton [m² mg Chl a⁻¹]; Φ_Cmax: Maximum quantum yield for carbon fixation [mol C mol photons⁻¹]; PP: Primary productivity [mg C m⁻³ day⁻¹].

Table 5.

Spearman-rank correlation coefficients (ρ) between photosynthetic parameters

	Chl a	Fv/Fm	αB	βB	PBmax	Ek	ā*ph	ΦCmax	PP
SST	−0.01	−0.25	0.11	0.23	0.75**	0.40	−0.27	−0.66*	−0.07
SSS	−0.44	−0.40	−0.07	−0.16	0.45	0.31	−0.63*	−0.65*	−0.32
NO3	−0.61*	−0.41	−0.18	−0.36	0.18	0.50	−0.66*	−0.87***	−0.63*
NO2	−0.42	−0.59*	−0.22	−0.52	0.21	0.23	−0.64*	−0.72**	–0.41
NH4	−0.60*	−0.04	−0.19	−0.12	−0.10	0.64*	−0.58*	−0.76**	−0.87***
PO4	−0.61*	−0.41	−0.18	−0.36	0.18	0.50	−0.66*	−0.87***	−0.63*
SiO2	−0.36	−0.55	−0.41	−0.68*	−0.18	0.24	−0.50	−0.58*	–0.34
MLD	−0.14	−0.31	−0.66*	−0.55	−0.54	−0.01	−0.16	−0.17	–0.24
Zeu	−0.54	−0.03	−0.02	0.09	0.19	0.73**	−0.71**	−0.77**	−0.66*
DFe	0.43	−0.10	−0.47	−0.26	−0.38	−0.31	0.37	0.35	0.27

Bold numbers indicate significant relationship between given parameters. Significance levels are denoted by numbers of asterisks:

^*= 0.05,

^**= 0.01,

^***= 0.001, n = 12.

SST: Sea surface temperature; SSS: Sea surface salinity; NO₃, NO₂, NH₄, PO₄, and SiO₂: Concentration of nitrate, nitrite, ammonium, phosphate, and silicate; MLD: Mixed layer depth; Z_eu: Euphotic layer depth; DFe: Dissolved iron concentration

Photophysiological parameters and their units were the same as shown in the caption of Table 4.

3.7. Primary productivity

Station D9 in the Sea of Okhotsk had a remarkably high PP value (326 mg C m⁻³ d⁻¹; Fig. 6G), followed by station B5 in the NS water. The Offshore stations had relatively low values of ~20 mg C m⁻³ d⁻¹, although station A5 showed a relatively high value among these stations (Fig. 6G). Significant relationships were found with other photosynthetic parameters: positive relationships with chl a, ā*_ph, or Φ_Cmax (p<0.05, Table 4) and a negative relationship with E_k (p<0.05, Table 4). A negative correlation between PP and Z_eu was also observed (p<0.05, Table 5).

4. Discussion

4.1. Phytoplankton abundance and biogeochemical properties

The Okhotsk station D9 showed a remarkably high chl a biomass (7.6 mg m⁻³) with sufficient macronutrient and DFe concentrations (Fig. 3A and Table 2), which suggests that this station had sufficient supplies of macro- and micronutrients to the phytoplankton growth. Although phytoplankton at the station D9 was in a bloom phase, the high DFe:NO₃ ratio of 0.349 (Fig. 2 and Table S1) suggests that nitrate could be depleted earlier than DFe at the station based on the nutrient consumption criterion of Nishioka et al. [2011] with the DFe:NO₃ ratio of 0.043 in the study area during spring and summer. According to Nishioka et al. [2011], higher and lower values than 0.043 in DFe:NO₃ ratio indicate the expected occurrence of nitrate and Fe limitation, respectively.

The SS waters had higher nitrate and DFe concentrations and showed similar SiO₄ and DFe concentrations among the stations (Table 2). However, chl a biomass at these stations were quite variable (0.4–6.2 mg m⁻³) (Fig. 3A). Despite the large differences in chl a biomass, the similar and low DFe:NO₃ ratios (0.012−0.013 nM/μM) (Fig. 2) suggest that Fe would potentially be the limiting nutrient based on the nutrient diagnosis of Nishioka et al. [2011] mentioned above.

The NS stations showed moderate to high chl a biomass (ca.1−2.6 mg chl a m⁻³) (Fig. 3A). All NS stations showed silicate concentrations one order of magnitude lower (<3.6 μM) than the values at the other stations (Table 2), which indicates that silicate was commonly the limiting nutrient for diatoms. According to Sarthou et al. (2005), the half-saturation constant values of silicate for the growth of diatoms are <3.9 μM. The silicate deficiency could be crucial in Oyashio waters because diatoms primarily contribute to spring blooms and the most significant phytoplankton group in the WSP [Mochizuki et al., 2002; Hattori-Saito et al., 2010; Suzuki et al., 2011; Nosaka et al., 2017; Yoshida et al., 2018]. It should be noted that nitrate was also depleted at station PS1. Nitrate could be another limiting nutrient based on the high DFe/NO₃ values in the NS waters (>0.040 nm/μm) (Fig. 2) except at station B5. However, it was commonly observed that silicate was first depleted by diatoms in the NS waters.

The Offshore stations, on the other hand, had low chl a biomass (0.5–1.1 mg m⁻³) and lower DFe values (Fig. 3A and Table 2), which suggests that Fe was the limiting nutrient. Station C3 was an exception, with a high DFe concentration (Table 2). Although this station might be in a different bloom phase and had different Fe sources, Fe would be the limiting nutrient when the nutrient diagnosis of Nishioka et al. [2011] was applied (0.010 nM/μM) (Fig. 2).

Overall, Fe was the principal or potential limiting nutrient in the Offshore waters. This supports the HNLC waters normally occupied in the WSP in summer (e.g., Tsuda et al., 2003; Suzuki et al., 2009; Fujiki et al., 2014). The depletion and relative deficits of Fe were consistent with previous Fe studies near the Bussol’ Strait [Yoshimura et al., 2010; Sugie et al., 2013; Suzuki et al., 2014].

4.2. Community composition of phytoplankton

The NS waters, where silicate could be the limiting nutrient for diatoms, were dominated by microphytoplankton except at station B6 (Fig. 3B), and the chain-forming diatoms such as C. debilis and C. furcellatus were predominant, particularly at station PS5 in the NS water (Table 3). Timmermans et al. [2001] indicated that Chaetoceros can outcompete other diatom genera under Fe-replete conditions. This notion has also been supported by other field studies [Hutchins et al., 1998; Tsuda et al., 2003; Marchetti et al., 2006].

The two stations at A2 and OP3 in SS waters showed the contrastive algal community composition as well as chl a biomass (Figs. 2A and 4; discussed in section 4.1). The exclusive dominance of diatoms with a high contribution of Chaetoceros at station A2 was similar to that in adjacent waters (e.g., station D9 and A5; Fig. 4), whereas there was a low diatom contribution at station OP3 (79.1%), with a considerable contribution of the silicoflagellates Dictyocha speculum (Table 3). Interestingly, most of Chaetocetos at station OP3 was present as resting spores (83.7%) (Tables 3 and S4). The high abundance of resting spores (i.e., cell dormancy) of Chaetoceros might allow the predominance of silicoflagellates and other diatoms (e.g., Fragilariopsis and Pseudo-nitzschia) instead of vegetative Chaetoceros cells at station OP3, similar to the offshore station A6 (Fig. 4).

In the Offshore waters, picophytoplankton was mainly predominant (Fig. 3B). The predominance of smaller phytoplankton in the low-Fe Offshore waters was advantageous in nutrient uptake because smaller cells have a higher surface-to-volume ratio (i.e., the allometry of a cell) [Taguchi, 1976; Sunda and Huntsman, 1997] (Fig. 3B). The SEM also revealed the frequent appearance of the small diatoms Fragilariopsis, Pseudo-nitzschia, and Thalassiosira oceanica at stations C3, B3 and A6 in the Offshore (Tables 3 and S4). Fragilariopsis is often observed in Fe-limited waters in the Southern Ocean [de Salas et al., 2011; Eriksen et al., 2018]. Mock et al. [2017] sequenced the genome of Fragilariopsis cylindrus and revealed that this psychrophilic species is tolerant to low Fe availability, which supports the dominance of Fragilariopsis in the Fe-limited Southern Ocean, although the ability of Fe acquisition may change among species [Wright et al., 2010]. Regarding Pseudo-nitzschia, Marchetti et al. [2012] indicated that this genus is highly plastic in Fe requirement and utilization efficiency using the iron storage protein ferritin [Marchetti et al., 2009]. The small diatom T. oceanica is also tolerant of low Fe availability [Sunda and Huntsman, 1995]. Although the diatom community at station C5 was dominated by Chaetoceros, note that small parmales comparably contributed to the phytoplankton community (Table 3). Parmales are small silicifying bolidophytes (< 2 μm) [Booth and Marchant, 1987; Ichinomiya et al., 2011]. The abundant presence of parmales was reported in the subarctic Pacific and near the Aleutian Islands [Konno et al., 2007]. Station C5 had the halved nitrate concentrations but with a comparable silicate concentration compared with the other offshore waters (Table 2), which might support the growth of the small silicifying parmales at the station. However, station A5 showed the high abundance of diatoms mainly consisting of Chaetoceros. The cluster analysis indicated that the community composition at station A5 was more similar to the coastal communities at stations D9 and A2, whereas station A6 showed a similar community composition to that of station OP3 in the SS waters (Fig. 4), where a low DFe concentration was observed even in the coastal region (Table 2). These heterogeneous distributions of DFe and diatoms might relate to complex physical properties around the Kuril Islands such as intense vertical tidal mixing [Ono et al., 2007; Yagi and Yasuda, 2012] and anticyclonic eddies that occur in summer along the Kuril-Kamchatka Trench [Yasuda et al., 2000; Kaneko et al., 2015].

The RDA results revealed that the community composition of phytoplankton in the studied area was mainly characterized by the two major diatoms Chaetoceros and Fragilariopsis (Fig. 5). This demonstrates clear ecological compartmentalization between these two diatoms, partially due to the differences in the responses to Fe availability and its requirement between these species (discussed above). This notion is supported by the opposite directions of Fragilariopsis to the DFe vector. The good agreement between the vectors of Cylindrotheca and NH₄ (Fig. 5) suggests that elevated NH₄ concentration could stimulate the growth of Cylindrotheca. Grant et al. [1967] and Underwood and Provot [2000] indicated that this genus shows optimal growth under higher NH₄ concentrations compared with other diatoms. This suggests rapid responses of this genus to NH₄ availability, which leads to the robust relationship between NH₄ concentrations and contribution of Cylindrotheca (Fig. 5). The RDA also suggested a coupling between DFe and Neodenticula seminae, which possessed heavily silicified frustules as observed by SEM; this indicates that Fe addition could enhance their growth. This species is quite important in the WSP because of their considerable sinking flux [e.g., Katsuki and Takahashi, 2005; Onodera et al., 2005; Shimada et al., 2006]. Therefore, this diatom species would play a crucial role in the biological carbon pump and a paleoceanographic proxy of primary production [Okazaki et al., 2005]. However, the physiology of this species has not been well documented thus far. The DFe and N.seminae vectors in Fig. 5 had similar directions; therefore, to our knowledge, this study was the first to indicate Fe availability controlled the abundance of N. seminae.

4.3. Photophysiological state of phytoplankton

Photosynthetic parameters did not show any significant relationship with DFe concentrations (Table 5). This would be partly due to the rapid utilization and uptake of nutrients including Fe [Cullen et al., 1992; Lomas et al., 2019] and the use of other Fe sources such as particulate Fe [Sugie et al., 2013]. The highest PP in the Okhotsk water at station D9 could be due to sufficient supply of nutrients, including DFe, in the Sea of Okhotsk (Fig. 6G and Table 2). The difference in PP between the two SS stations could be partly caused by the abundant resting cells of Chaetocerous at station OP3 (Table 3; discussed in section 4.2 above). Values of α^B and P^B_max were overwhelmingly high at stations PS1 and PS5 in the NS (Fig. 6A, C), although the chl a biomass was relatively low (~0.8 mg m⁻³) at these stations (Fig. 3A). This discrepancy implies the significant grazing and/or sinking effects on the biomass. Indeed, high abundance of breakdown products of chlorophyll a (i.e., phaeopigments) were observed at these stations [K. Suzuki, personal communication].

Values of PP were negatively correlated with Z_eu and E_k and positively correlated with chl a (Tables 4 and 5), which indicates that phytoplankton communities in the WSP were well-acclimated to in situ light conditions and that their standing stock controlled PP in the WSP regardless of the water masses. The photosynthetic parameters showed a parallel change between P^B_max and α^B (i.e., the positive and significant relationship between these parameters) (Table 4), the so-called E_k-independent variability [Behrenfeld et al., 2004]. Although E_k-dependent variability has been well studied both in the field and in the laboratory and noted as a light-shade acclimation [e.g., Falkowski and Owens, 1980; Sakshaug et al., 1997; MacIntyre et al., 2002], Behrenfeld et al. [2004; 2008] pointed out that E_k-independent variability was driven by downstream photochemical electron transport such as the utilization ratio of NADPH and ATP. Diatoms possess capabilities of exchanging ATP between plastids and mitochondria and of changing the NADPH/ATP ratio [Bailleul et al., 2015]. The diatom blooms and diatom-dominated community sometimes showed E_k-independent variability [e.g., Platt et al., 1992; Yoshie et al., 2010; Isada et al., 2013; Morán and Scharek, 2015]. Yoshie et al. [2010] suggests that the level of ammonia compared with nitrate could contribute E_k-independent variability because the difference in the redox state of these nitrogen compounds could affect the quota of photosynthetic energy for carbon fixation and nitrogen assimilation. The positive relationship between E_k and ammonium concentrations might reflect changes in metabolic activities downstream of the photochemical electron transport chain (Table 5). SST was significantly and positively correlated with P^B_max (Table 5), and temperature could thus be an important driver of photosynthesis during the expedition because temperature controls the activity of the carbon fixation enzyme ribulose-1,5-bisphosphate carboxylase/oxygenase (RuBisCO) [e.g., Li et al., 1984; Raven and Geider, 1988; Young et al., 2015; Yoshida et al., 2018]. Interestingly, E_k was significantly and positively correlated with Z_eu (Table 5), which suggests that phytoplankton was well-acclimated to the light environment (i.e., light-shade acclimation). The negative relationship between α^B and MLD (Table 5) also supported the light-shade acclimation because a smaller ratio of Z_eu/MLD (i.e., shallower light penetration) led to a lower E_k and a higher α^B, which indicates low-light acclimation, and vice versa.

5. Conclusions

The community composition and photosynthetic physiology of phytoplankton were examined in less-studied shelf waters in the WSP as well as the oceanic region during summer. Diatom assemblages in the shelf and offshore waters were generally dominated by the centric Chaetoceros and the mixture of centric and pennate species, respectively. Based on the DFe:NO₃ ratios (Fig. 2), the offshore waters were limited by Fe availability. Shelf waters could be limited by nitrate and/or silicate; however, we infer silicate first limited the growth of diatoms in the NS waters. In addition, the phytoplankton communities in the NS waters were distinct from those in the other waters (Fig. 4); this finding indicates that the NS waters upstream of Oyashio had different nutrition status from other southern and offshore waters, which led to the different community compositions. Although F_v/F_m did not show any correlation with environmental parameters except nitrite (Table 5), PP was significantly correlated with the concentrations of chl a (Table 4), which indicates that biomass was a major controlling factor of PP. For the first time, our RDA revealed that Fe availability could control the abundance of Neodenticula seminae (Fig. 5), which had heavily silicified frustules. Because N. seminae was the most abundant diatom species collected with a moored sediment trap in the WSP [Onodera et al., 2005], this species could play an important role in the biological carbon pump in the study area. Recently, Tréguer et al. [2018] also suggested that large and heavily silicified species have more potential to contribute to carbon sinks via the biological carbon pump.

Key Points:

• Diatoms in the shelf and offshore waters were mainly composed of the centric Chaetoceros and a mixture of centric and pennate species.

• Macronutrients and dissolved Fe (DFe) could have regulated the phytoplankton composition in the shelf and offshore waters, respectively.

• The pennate diatom Neodenticula seminae, which has often been observed in sinking particles, was positively correlated with DFe levels.