Effects of Ocean Currents in the Western Pacific Ocean on Net-Phytoplankton Community Compositions

Phytoplankton are known as important harbingers of climate change in aquatic ecosystems. This study investigated phytoplankton community structure in the western Pacific Ocean (WPO) in 2017 and revealed the spatial variability of phytoplankton in community composition and abundance, as well as their relationship to physical processes and environmental factors. The phytoplankton community was mainly composed of Dinophyta (221), followed by Bacillariophyta (105), Cyanophyta (4), and Chrysophyta (2). The cyanobacteria Trichodesmium were the dominants throughout the study period. Correlation analysis showed that dinoflagellates were mainly affected by temperature, while diatoms were significantly correlated with nutrients (silicate, phosphate, nitrite, nitrate). Phytoplankton was divided into five groups by cluster analysis, and the distribution of different groups was related to circulation and hydrological characteristics. In contrast, the highest abundance of diatoms and dinoflagellates was found in the New Guinea Coastal Current (NGCC) region, while the highest abundance of cyanobacteria was found in the Northern Equatorial Counter Current (NECC) region. Overall, we found that not only temperature and salinity, but also ocean currents and nutrients, influence the distribution of phytoplankton communities in the WPO.


Introduction
As the most important producers in marine ecosystems, phytoplankton not only absorb carbon dioxide and regulate global climate change, but also provide energy through photosynthesis [1]. Phytoplankton communities play an important role in biogeochemical cycles and pelagic food webs, and they also support the energy needs of marine ecosystems [2]. They are widely distributed in the upper layer, and the distribution of phytoplankton species and their community structure are usually associated with the dominant marine environment (unique nutrient structure) [3]. The distribution of phytoplankton in seawater is usually strongly influenced by ocean circulation and mesoscale hydrographic features [4]. Some specific currents and water masses are inhabited by specific native species, a property that can be used to indicate water movement [5]. For this reason, the distribution of many phytoplankton species is closely related to unique environmental factors in many marine ecosystems.
The Pacific Ocean covers about a third of the earth and nearly half of the sea surface around the world. The boundaries of the tropical western Pacific Ocean (WPO) are very

Material and Method
The cruise was carried out from 142-163 • E and 1-40 • N in the WPO onboard R/V "Dongfanghong 2", covering 63 stations from October 6th to December 7th in 2017. As shown in Figure 1, the survey area was divided into two sections: Section A (stations assigned to the longitude of 143-146 • E) and Section B (stations assigned along the equator). These stations were sampled between 0 and 200 m depth by vertical tows with the plankton net (mouth opening 0.25 m 2 , mesh size 20 µm). Samples were fixed in 5% buffered formalin and stored in the dark. In the laboratory, phytoplankton samples are observed under a microscope (AE2000, Motic, Xiamen, China) [15]. Phytoplankton identification was conducted as described by Jin [16], Isamu Y [17], and Sun [18]. Species identification was as close as possible to the species level.
Seawater samples were collected by a SeaBird CTD (SBE 9/11 plus) equipped with Go-Flo bottles, and temperature and salinity were recorded at the same time. The temperature and salinity of the water column were calculated by means of a trapezoidal integration of the different levels of seawater temperature and salinity. Nutrient samples from different layers were placed in PE bottles and stored at −20 • C for laboratory nutrient analysis [19]. Nitrate, nitrite, ammonium, phosphate, and silicate were also analyzed on board by spectrometric methods [20]. Chlorophyll a (chl a) samples were filtered through Whatman GF/F filters (0.7 µm) for seawater samples (1 L) and subsequently saved under -20 • C. Chl a was extracted in laboratory without light by 90% acetone, and then Turner fluorimeter (model 10-AU) was used to measure the chl a concentration [21]. The sampling layers were 5 m, 30 m, 75 m, 100 m, 150 m, and 200 m. Dominance index (Y) was calculated to describe the species dominance in the phytoplankton community. The calculation equation was as follows: where ni is the number of the individual species, N is the total number of all species, and fi is the occurrence frequency of the species in a sample. The abundance of phytoplankton cells in water column was calculated through the trapezoidal integral method [24]: where P is the average value of phytoplankton abundance in water column, Pi is the abundance value of phytoplankton in layer i, i + 1 is the layer i + 1, Dn is the maximum sampling depth, Di is the depth of layer i, and n is the sampling level.
Horizontal and depth-integrated distribution of phytoplankton and physiochemical parameters were projected using Ocean Data View 4.7.6 and ArcGIS 10.8. The histogram was plotted with Origin (Version 8.5). Pearson correlation and canonical correspondence analysis (CCA) between assemblages and physicochemical parameters were performed using the R package vegan (version 2.5-7) [25] to explain the relationship between the environmental parameters (temperature, salinity, and nutrients) and phytoplankton community structure. Dominance index (Y) was calculated to describe the species dominance in the phytoplankton community. The calculation equation was as follows: where n i is the number of the individual species, N is the total number of all species, and f i is the occurrence frequency of the species in a sample. The abundance of phytoplankton cells in water column was calculated through the trapezoidal integral method [24]: where P is the average value of phytoplankton abundance in water column, Pi is the abundance value of phytoplankton in layer i, i + 1 is the layer i + 1, D n is the maximum sampling depth, D i is the depth of layer i, and n is the sampling level. Horizontal and depth-integrated distribution of phytoplankton and physiochemical parameters were projected using Ocean Data View 4.7.6 and ArcGIS 10.8. The histogram was plotted with Origin (Version 8.5). Pearson correlation and canonical correspondence analysis (CCA) between assemblages and physicochemical parameters were performed using the R package vegan (version 2.5-7) [25] to explain the relationship between the environmental parameters (temperature, salinity, and nutrients) and phytoplankton community structure.

Phytoplankton Species Composition
Phytoplankton samples from the WPO were analyzed, and 332 species belonging to 68 genera in 4 phyla were identified, including Bacillariophyta, Dinophyta, Cyanophyta, and Chrysophyta. There were 105 diatoms belonging to 37 genera (31.63% of total species), 221 species in 28 genera of dinoflagellates (66.57%), 4 cyanobacteria species, and 2 chrysophyceae species ( Table 1). The distribution of phytoplankton ( Figure 2a) is mainly determined by Cyanobacteria. The distribution of dinoflagellate ( Figure 2b) shows that most of the sites of dinoflagellate are evenly distributed, and diatom is more dominant in some stations near the shore. Trichodesmium thiebautii is the main species of Trichodesmium and is abundantly distributed in the surveyed sea area (Figure 2c). The abundance of Trichodesmium reaches the maximum value (12.905 × 10 3 cells m −3 ) at the A35 station of the investigated sea area. The distribution of dominant species (Figure 2d) shows that there is a rich diversity of phytoplankton in the investigated sea area. The dominant species of dinoflagellate in this survey is Ceratium kofoidii, and the maximum value appears at A28 station (5.619 × 10 3 cells m −3 ). Planktoniella foromsa is the dominant species of diatom in this survey, and the maximum value appears at B19 station (13.276 × 10 3 cells m −3 ). Thalassiothrix longissima is a common species of Kuroshio, which is dominant at A01 and A02 stations, and reaches the peak at B18 station with higher diatom abundance ( Table 2). Cluster analysis was used to analyze the similarity of phytoplankton taxa. The results of the cluster analysis were used to analyze the similarity between phytoplankton taxa based on their abundance during the 2017 cruise. The cluster analysis revealed five distinct phytoplankton communities ( Figure 3). Phytoplankton groups containing fewer than four stations were not considered as significant clusters. Clusters were assigned different color symbols and plotted in the sampling area ( Figure 3).

Environmental Characteristics of the Survey Area
Phytoplankton clustering analysis divided the 63 stations surveyed into five gro ( Figure 3). The different groups had different hydrological characteristics. T hydrological characteristics included temperature (T), salinity (S), chl a, silicate (SiO phosphate (PO4 3− ), nitrate (NO3 − ), nitrite (NO2 − ), and ammonium (NH4 + ). We found the five groups had different characteristics. Group A had the highest mean temperat the highest mean salinity, the highest mean nutrient concentration, and a deeper m layer with a thickness of about 125 m; Group B had the shallowest mixed layer wi thickness of about 30 m; Group C had the highest mean silicate concentration; Grou had the lowest mean temperature and the lowest mean salinity; and Group E moderate mean temperature, salinity, and nutrient levels (Table 3, Figure 4).

Environmental Characteristics of the Survey Area
Phytoplankton clustering analysis divided the 63 stations surveyed into five groups ( Figure 3). The different groups had different hydrological characteristics. These hydrological characteristics included temperature (T), salinity (S), chl a, silicate (SiO 3 2− ), phosphate (PO 4 3− ), nitrate (NO 3 − ), nitrite (NO 2 − ), and ammonium (NH 4 + ). We found that the five groups had different characteristics. Group A had the highest mean temperature, the highest mean salinity, the highest mean nutrient concentration, and a deeper mixed layer with a thickness of about 125 m; Group B had the shallowest mixed layer with a thickness of about 30 m; Group C had the highest mean silicate concentration; Group D had the lowest mean temperature and the lowest mean salinity; and Group E had moderate mean temperature, salinity, and nutrient levels (Table 3, Figure 4). Table 3. Average (± standard deviations) values for nutrients (µmol L −1 ), temperature ( • C), salinity, and chlorophyll a (µg L −1 ) for each phytoplankton community group identified by the cluster analysis in the WPO.

Phytoplankton Community Structure of Five Groups
Cluster analysis showed that phytoplankton communities have a spatial distribution structure in the currents of the WPO. In the five groups, the phytoplankton had different proportions of cell abundance. The cell abundance of the four-phyla phytoplankton also differed in the five groups ( Figure 5). The dominant species of phytoplankton had different mean cell abundance in the five groups (Table 4). Overall, the highest abundance of diatoms and dinoflagellates was in Group A, and the highest abundance of cyanobacteria in Group C.
Group A was distributed along the equator and mainly affected by the NGCC. The total phytoplankton abundance in Group A was not the highest (2.12 × 10 5 cells m −3 ) among the five phytoplankton groups, but the abundance of diatoms and dinoflagellates in Group A was the highest (1.22 × 10 4 cells m −3 and 0.98 × 10 4 cells m −3 ). Cyanobacteria, diatoms, dinoflagellates, and chrysophyceaes accounted for 89.49%, 5.75%, 4.64%, and 0.12% of the total phytoplankton, respectively. The abundance of P. foromsa (2.39 × 10 3 cells m −3 ) and T. longissima (9.88 × 10 2 cells m −3 ) was higher than other groups. Additionally, R. intracellularis was not found in Group A.

Phytoplankton Community Structure of Five Groups
Cluster analysis showed that phytoplankton communities have a spatial distribution structure in the currents of the WPO. In the five groups, the phytoplankton had different proportions of cell abundance. The cell abundance of the four-phyla phytoplankton also differed in the five groups ( Figure 5). The dominant species of phytoplankton had different mean cell abundance in the five groups (Table 4). Overall, the highest abundance of diatoms and dinoflagellates was in Group A, and the highest abundance of cyanobacteria in Group C. abundance (6.78 × 10 2 cells m −3 ) in Group D among the five groups. Group E, located in the middle of the sampling area, was mainly affected by the NEC. Group E was the group with the lowest total phytoplankton abundance (7.71 × 10 4 cells m −3 ), and the abundance (4.51 × 10 3 cells m −3 , 0.01 × 10 3 cells m −3 , 0.07 × 10 3 cells m −3 ) of T. hildebrandtii, P. foromsa, and T. longissima in this group was also very low. Cyanobacteria, diatoms, dinoflagellates and chrysophyceaes accounted for 87.10%, 5.21%, 7.56%, and 0.13% of the total phytoplankton, respectively.  Group A was distributed along the equator and mainly affected by the NGCC. The total phytoplankton abundance in Group A was not the highest (2.12 × 10 5 cells m −3 ) among the five phytoplankton groups, but the abundance of diatoms and dinoflagellates in Group A was the highest (1.22 × 10 4 cells m −3 and 0.98 × 10 4 cells m −3 ). Cyanobacteria, diatoms, dinoflagellates, and chrysophyceaes accounted for 89.49%, 5.75%, 4.64%, and 0.12% of the total phytoplankton, respectively. The abundance of P. foromsa (2.39 × 10 3 cells m −3 ) and T. longissima (9.88 × 10 2 cells m −3 ) was higher than other groups. Additionally, R. intracellularis was not found in Group A.
Group C was mainly affected by the NECC. Group C had the highest abundance of total phytoplankton (1.45 × 10 6 cells m −3 ). Group C of phytoplankton was mainly composed of cyanobacteria; cyanobacteria accounted for 98.99% of the total phytoplankton, diatoms accounted for 0.46% of the total phytoplankton, dinoflagellates accounted for 0.50% of the total phytoplankton, and chrysophyceaes accounted for 0.05% of the total phytoplankton. C. marginato-lineatus had the highest abundance (7.19 × 10 2 cells m −3 ) among the five groups in Group C.
Group D mainly contained stations affected by the STCC and KC. Cyanobacteria, diatoms, dinoflagellates and chrysophyceaes accounted for 91.71%, 4.63%, 3.29% and 0.37% of the total phytoplankton respectively. Fragilariopsis doliolus had the highest abundance (6.78 × 10 2 cells m −3 ) in Group D among the five groups.

Phytoplankton Distribution in Relation to Environmental Factors
The influence of environmental factors on the phytoplankton community structure in the WPO was assessed using Pearson s correlation ( Figure 6) and CCA analysis (Figure 7). The phytoplankton community in the region was significantly influenced by environmental factors. Phytoplankton cell abundance was extremely significantly correlated with cyanobacteria (p < 0.001), indicating that cyanobacteria are the main component of phytoplankton in this survey area. Diatom was extremely significantly correlated with phosphate and nitrate (p < 0.01) and was also significantly correlated with silicate and nitrite (p < 0.05). This shows that the abundance of diatom is more affected by nutrients. There was a significant correlation between dinoflagellate and temperature (p < 0.05), and temperature plays an important role in the growth of dinoflagellate. Different dominant species had different responses to the aquatic environment. The abundance of T. erythraeum, P. compressum, C. marginato-linetus, and P. Leniculatum were positively correlated with nitrates and silicates concentration. The abundance of P. foromsa, T. hildebrandtii were positively correlated with phosphate and nitrite concentration.

Hydrological Conditions and Corresponding Phytoplankton Community Structure
The study spanned five different hydrological characteristics distributed across the 40 • N to equatorial cross-section, Kuroshio region, subtropical gyre, transitional zone, warm pool, and equator region [26][27][28][29]. The difference in salinity between the surface water in the transition zone and the Subsurface Chlorophyll Maximum (SCM) confirms the existence of positive precipitation budget phenomenon in this area [30]. However, at higher resolution, neither phytoplankton species richness nor species distribution can be strictly distinguished by hydrological characteristics (Figures 4 and 5). On the contrary, according to the results obtained in this study, phytoplankton species composition and abundance tend to change gradually along the cross-section, resulting in some transition zones between different hydrological characteristics. For example, from the Kuroshio area to the transition area, nutrient conditions gradually changed to oligotrophic conditions, and correspondingly, the phytoplankton abundance gradually decreased [31]. The composition and abundance of phytoplankton community also changed spatially. Although some mesoscale vortexes and secondary mesoscale gyres may cause instability in local water bodies, this spatial distribution of diatoms occurs when nutrient gradients are kept changing along latitudes for a certain period. However, phytoplankton gradual adaptation to oligotrophic conditions depends on the species composition of diatoms and dinoflagellates [32][33][34].
In addition to changes in phytoplankton abundance between different hydrological characteristics, variations in diatoms and dinoflagellates abundance within the same hydrological characteristics were also found (Figures 4 and 5), suggesting that mesoscale circulation could play an important role in phytoplankton distribution [35]. Due to their poor activity and high potential growth rate, diatoms can reproduce rapidly in circulation and in water with high nutrient content. However, the disadvantage of poor mobility comes with its advantage. The circulation can not only bring new nutrient supplements, but also enable diatoms to distribute evenly in the sea area with strong circulation [36]. This also explains the high diatom abundance in Group A under the influence of NGCC. In fact, dinoflagellates are more susceptible to circulations and vortices, and the more violent the circulations and vortices are, the more their growth is inhibited [37]. The effects of circulations and vortices on dinoflagellates include inhibition of cell division, destruction of cell morphology, and inhibition of nutrient transport [38]. The characteristics of high temperature, high salinity, and high flow rate in the Kuroshio area exactly inhibited the abundance of dinoflagellates in this area, which was consistent with the previous research results [39].

Distribution of Trichodesmium
In the present study, Trichodesmium was the dominant cyanobacteria species. Marine Trichodesmium has been considered the most critical autotrophic nitrogen-fixing cyanobacteria since the 1960s [40]. Trichodesmium can be divided into two forms: clusters and free filaments. Trichodesmium is suitable for living in waters above 20 • C and has a special cellular air sac structure that allows it to move vertically within the upper 100 m of the ocean water column [41]. In the process of water bloom formation by Trichodesmium, a large amount of nitrogen is often fixed in a relatively short period of time. Therefore, the study of the nitrogen fixation rate of Trichodesmium is crucial for estimating the rate of nitrogen fixation in the ocean [42]. Trichodesmium is an important nitrogen-fixing organism in the ocean and a major contributor to the new productivity of the oligotrophic sea area [43]. The abundance of Trichodesmium in the tropical oligotrophic sea area is an issue of great concern [44].
In this study, cyanobacteria bloom was observed near New Guinea. The discovery of high abundance of Trichodesmium near the coast is consistent with the results of Campbell [45]. However, until now, it has not been clear how environmental factors control the latitude distribution of Trichodesmium. According to our results, temperature has an impact on the spatial distribution of Trichodesmium, which is similar to previous observations of tropical and subtropical oceans [46,47]. The optimum temperature for the growth and nitrogen fixation of Trichodesmium was between 20 • C and 30 • C. Our results show that there was a positive correlation between temperature and Trichodesmium abundance (Figure 6).
At present, there have been reports on the abundances of other oligotrophic salts in the oceanic region: Bonnet found that Trichodesmium has the highest abundance of trichomes L −1 (1.85) in the equatorial WPO [48]; Zhang [49] found that the average abundances of Trichodesmium in the central, eastern, and southern Indian Ocean were 1.76, 0.87 and 1.52, respectively. Therefore, this study investigates the high abundance of Trichodesmium, which is consistent with previous studies. Previous studies have not clarified which factors are the main causes of Trichodesmium growth (possibly temperature, wind, iron, phosphorus, etc.) [50,51]. Many researchers believe that temperature is the most important factor affecting the growth of Trichodesmium [52]. However, we believe that there is no single positive correlation between temperature and Trichodesmium growth, which is consistent with the study by Chang [50]. In the tropical WPO, where the temperature was not restricted, Group A had the highest temperature, but the abundance of Trichodesmium in Group A was not the highest. The highest value of Trichodesmium was in Group C, which had the second-highest temperature (Table 3, Figure 5). The cluster analysis divided the five groups according to the abundance of phytoplankton, which was consistent with the currents (Figure 3). We believe that marine physical processes such as circulation and hydrological characteristics have a profound impact on the spatial distribution of phytoplankton in the WPO.

Dominant Species and Their Preferred Environmental Factors
The conditions suitable for phytoplankton are often different from one community to another and even from one species to another. Comparing the dominant phytoplankton species of five groups, we conclude that P. leniculatum was the dominant species in the study area. P. leniculatum belongs to Dinophyta, and it is widely distributed in the world, including the Pacific Ocean, the Indian Ocean, the waters near Madagascar, and the Andaman Sea [53]. In the study area, the abundance of P. leniculatum in Group B was the lowest, which was affected by STCC, and it had the lowest average nutrient concentration. CCA analysis showed that P. leniculatum in Group B had a positive correlation with nitrate and silicate. Different from others, C. kofoidii was distributed evenly and had a high abundance in the whole study area. C. kofoidii reached the highest abundance in Group B, which was critically affected by Pacific Subtropical gyre. Atmospheric nitrogen fixation in the ocean is an important source of new nitrogen in the surface waters, which stimulates phytoplankton productivity and provides fuel for biological pumps. Trichodesmium is the main group responsible for marine nitrogen fixation in tropical waters [42]. The R. intracellularis has been shown to provide significant nitrogen input to the ocean on a regional scale [54]. The results of this study show that R. intracellularis was not observed in the equatorial region, while there was a high abundance in the region of transitional zone and the Kuroshio region. CCA analysis showed that R. intracellularis prefers to bloom in areas with higher nitrate and silicate.

Conclusions
This study investigated phytoplankton community structure in the WPO in 2017 and revealed the spatial variability of phytoplankton in community composition and abundance, as well as their relationship with physical ocean processes and environmental factors. A total of 332 species of phytoplankton were identified in this survey. The highest abundance of phytoplankton was found in the NECC and equatorial regions. Trichodesmium was widely distributed in the study area and reached the peak in WPWP, dinoflagellates were mainly affected by temperature, and diatoms were significantly correlated with nutrients (phosphate, nitrate, silicate, and nitrite). Phytoplankton were divided into five groups by cluster analysis, and the distribution of different groups was related to circulation and hydrological characteristics. These results show that physical ocean processes such as circulation and hydrological characteristics have a profound influence on the spatial distribution of phytoplankton. Different currents divide phytoplankton into different groups in space. In this investigation, we found that not only temperature and nutrient salinity, but also currents and water mass movements, affect the distribution of phytoplankton communities in the WPO. Despite the baseline data and information provided by this study, the phytoplankton of the WPO remain a mystery to us, especially the distribution of phytoplankton throughout the water column and eddies. Therefore, more long-term community studies are needed to further explore the role of phytoplankton in the marine biogeochemistry of the WPO.