Phytoplankton diversity and chemotaxonomy in contrasting North Pacific ecosystems

Expedition- location and time

A Sea to Space Particle Investigation cruise aboard the Schmidt Ocean Institute R/V Falkor was conducted from January 24 to February 20, 2017, in North Pacific (Fig. 1). The aim of the cruise was to connect the radiometric properties (ocean colour) as well as ecological mechanisms of carbon export (Durkin et al., 2022) with the trophic state of the ocean, and use those data to develop algorithms and phytoplankton proxies for the NASA’s PACE mission (pace.oceansciences.org).

Sampling

Sampling was done along the investigated transect at Station 1 (ST1) and Station 2 (ST2) in NPSG, and Station 3 (ST3) in CCS (Fig. 1). Each station represents a group of sampling sites (Table 1) where CTD (conductivity, temperature, depth) casts were deployed at three depths: the surface layer (S), deep chlorophyll maximum (DCM), and mixed layer depth (MLD), with exception at CTD 14 where additional sample was taken below mixed layer depth (BMLD) (Table 1). Due to the strong physical forcing, water column was well mixed as shown in CTD profiles of all three stations in Durkin et al. (2022) (Fig. S1). Samples for phytoplankton (n = 38) and pigment analyses (n = 38) were taken by 10 L Niskin rosette sampler equipped with CTD and other sensors. For more detailed taxonomic analyses, additional samples (n = 38) were taken from the same Niskin bottles, and volume of 400 mL seawater was filtered using weak vacuum onto 0.8 µm polycarbonate filters analyzed on SEM as described in (Šupraha, Ljubešić & Henderiks, 2018).

For qualitative plankton analysis, another set of samples (n = 27) was taken from the Niskin bottles and filtered through 20 µm mesh. Discrete phytoplankton and net phytoplankton samples were fixed with 2% neutralized formaldehyde and stored in 250 mL bottles until analyses in the Laboratory of Biological Oceanography, Department of Biology, University of Zagreb. Triplicate 4 L seawater samples were filtered on GF/F filters for phytoplankton pigment analysis and stored in liquid nitrogen until the high-performance liquid chromatography (HPLC) analysis in the NASA’s Goddard Space Flight Center, following methods described in Hooker et al. (2012).

Phytoplankton community analysis

Light microscopy (LM) was used to determine phytoplankton composition and abundance. Subsamples of 50 or 100 mL, depending on cell density, were settled for 24 h and 48 h respectively and analyzed under a Zeiss Axiovert 200 inverted microscope using the Utermöhl method (Utermöhl, 1958) as described in (Šupraha et al., 2016). For additional taxonomic analyses, net samples were analyzed with the Zeiss Axiovert 200 inverted microscope and images of all species were taken and analyzed with Zeiss AxioVision SE64 (version 4.9.1). Micrograph plate of dominant taxa was made and edited using Adobe’s Photoshop CC 2015 and Illustrator CC 2017.

Phytoplankton are comprised of a phylogenetically diverse group of both prokaryotic and eukaryotic organisms. Because of that, classification is much debated with different systematic grouping (Bray & Curtis, 2006; Roy et al., 2011; Thomas et al., 2012; Pal & Choudhury, 2014). Therefore, a simpler approach for classification will be presented in this paper with focus on morphological characteristics of most abundant forms only: cyanobacteria, diatoms, dinoflagellates, coccolithophores, cryptophytes and “others”–including phytoflagellates, silicoflagellates, ciliates and other genera. Phytoplankton were classified on size variation using the equivalent spherical diameter (ESD) of cells as nanophytoplankton (ESD 2–20 µm) and microphytoplankton (ESD 20–200 µm).

Trophic indices and spatial distribution

Pigment average concentrations were calculated in order to get F_p index using formula by Claustre (1994): F_p = (sum of average concentrations of fucoxanthin and peridinin)/(sum of average concentrations of all primary pigments). Spatial distribution across investigated transects was visualized by creating one chart showing abundances of phytoplankton groups, and another one with distribution of the subset of pigments that best correlate to phytoplankton community (the correlation test explained later in Statistical analysis section). Chart plotting and calculations were made using the software Grapher 12 (GoldenSoftware) and Microsoft Office 365 ProPlus (Microsoft Corporation, version 1705), respectively.

Satellite data

To illustrate distinct trophical ecosystem-specific properties of investigated area, monthly average satellite data (February 2017) of MODIS Aqua Sea Surface temperature and Chlorophyll a concentration, as well as Apparent Visible Wavelength (AVW) was used. AVW is an optical water type classification that allows for a single, highly sensitive metric to combine the information about the spectral shape of the ocean colour (Vandermeulen et al., 2020), where spectral shift in AVW indicates changes in oceanic components contribution to the ocean color. In open ocean, ocean colour is driven by the phytoplankton community and associated detrital component, and in coastal ocean dissolved organic matter and sediment can contribute as well. As such, it is a great geophysical tool to evaluate spatial and temporal changes across oceanic ecosystems.

Statistical analysis

Several statistical analyses were done using Primer 7.0. (2021; PRIMER-E, Ltd., Auckland, New Zealand) to test similarities between ST1, ST2 and ST3, and correlation between phytoplankton counts and pigment data in order to gain better understanding of community diversity.

North Pacific ecosystem and water column hydrography

Satellite data monthly averaged for February 2017 confirm that ST1, ST2 (NPSG), and ST3 (CCS) are located in distinct ecosystems (Fig. 1). NPSG has higher sea surface temperature compared to CCS (MODIS Aqua Sea Surface temperature data, Fig. 1A), where Chl a concentration is much higher (MODIS Aqua Chlorophyll a concentration data, Fig. 1B). Furthermore, MODIS Aqua Apparent Visible Wavelength data (Fig. 1C) indicates differences between two sampled environments based on the spectral shape of the color of the ocean: eutrophic CCS (higher AVW) and oligotrophic NPSG (lower AVW).

The deep chlorophyll maximum layer (DCM) was defined as highest fluorescence signal encountered during station profiles. For profiles collected at ST1 and ST2, it is set at ∼130 m, while it was found at much shallower depths at coastal ST3 (∼30 m). As expected, mixed-layer depth (MLD, calculated as the depth at which density differed from the mean density in the top 10 m by <0.05 kg m⁻³), was sitting in proximity of the DCM, at ∼130 m for ST1 and ST2, and at ∼90 m depth at ST3.

Phytoplankton diversity of North Pacific Ocean

The encountered phytoplankton community was mostly comprised of coccolithophores (35.5%), diatoms (25.2%) and dinoflagellates (19.5%) while cryptophytes, phytoflagellates and silicoflagellates, etc. were included in group “other” that makes 19.8% of phytoplankton counts. A total of 207 taxa have been determined from both Niskin and net samples of which: 106 diatoms, 48 coccolithophores, 41 dinoflagellates, seven other autotrophs, four heterotrophs, and one cyanobacterium. Cryptophytes were observed but were not identified to the genus level (Table S1). Of the 207 taxa, more than a half (113) taxa are found only in net samples: 42 diatoms, 40 coccolithophores, 27 dinoflagellates and four other heterotrophs.

Spatial distribution of phytoplankton groups using microscopy and pigments

Microscopy counts resulted in abundances of phytoplankton groups that indicate lower biomass of micro- and nanophytoplankton at NPSG oligotrophic ecosystem (ST1 and ST2) in comparison to eutrophic CCS (ST3). Moreover, results elucidate variable spatial distribution of microphytoplankton, while spatial distribution of nanophytoplankton is even (Figs. 2A and 2B). Diatoms of micro-fraction increased for an order of magnitude with the transition to ST3, while distribution of dinoflagellates, coccolitophores, and other phytoplankton groups of microphytoplankton stay constant across the investigated transect (Fig. 2A). Nano fraction of diatoms, dinoflagellates, and coccolitophores had even distribution across stations, while “other” cells (e.g., cryptophytes) exhibited similar behaviour to micro-scale diatoms, increasing their abundances at ST3 (Fig. 2B).

Average pigment concentrations encountered on transect (Table S4) show F_p index that is higher at ST3 (0,087), and lower at ST1 (0,018) and ST2 (0,021). Alloxanthin, zeaxanthin, divinyl chlorophyll b (DVChl b), and lutein are the pigment set with the highest correlation to phytoplankton counts, as identified by the BEST global test that resulted in Spearman correlation coefficient (Rho = 0.532) with p <0.1% significance level (Table S3). Spatial distribution of these four pigments and divinyl chlorophyll a (DVChl a) across stations elucidates two clearly distinct environments (Fig. 3A). ST1 and ST2 exhibited higher concentrations of DVChl a, and zeaxanthin, the biomarkers for Prochlorococcus and Synechococcus (respectively), implying the cyanobacteria domination in this region. Entering ST3, concentrations of previous pigments fall substantially, while concentrations of cryptophytic biomarker alloxanthin rise. Moreover, we observed higher concentrations of 19′-hexanoyloxyfucoxanthin (19HF) and fucoxanthin (Table S4), biomarkers for coccolithophores and mostly diatoms, respectively. Biomarkers peridinin and prasinoaxanthin also dominated at ST3, representing high abundances of dinoflagellates and prasinophytes, respectively (Table S4). Furthermore, there is a strong increase of total Chl a concentration at ST3, when compared to oligotrophic ST1 and ST2 (Fig. 3B).

Similarity between stations and dominant taxa

Pairwise test of ANOSIM analysis displayed significant differences in phytoplankton community abundance and composition between ST1 and ST3, and ST2 and ST3 with R-value being 0.579 and 0.612, respectively (Table S2). Taxa diversity and abundances were largest at eutrophic ST3, while oligotrophic ST1 and ST2 exhibited similar community structure.

Nano-scale dinoflagellates and coccolithophores contributed the most to dissimilarity of both ST1 and ST2 according to SIMPER analysis results (Table 2). Next were phytoflagellates with high contribution to the ST1 dissimilarity, while at ST2 that was Rhizosolenia hebetata f. semispina (Table 2). Nano-scale coccolitophorids (ESC 5–10 µm), cryptophytes and Pseudo-nitzschia pseudodelicatissima contributed the most to dissimilarity of ST3 (Table 2).

Dominant taxa by stations were defined as species and groups with abundance >10⁴ cells L⁻¹, and the frequency of occurrence in samples >50% (Table 3), and some of them are shown on micrographs (Fig. 4). Dominant taxa present along the whole transect, but reaching highest abundances at ST3 were cryptophytes, Gyrodinium spp, Nitzschia bicapitata, nano-dinoflagellates and nano-coccolithophores (ESC <10 µm) (Table 3). On the other hand, some dominant taxa were present at only one station. Species specific to ST1 were nano-scale Gyrodinium sp. (ESC <20 µm), Gymnodinium spp., Michaelsarsia adriaticus, N. braarudii, and Nitzschia sp. Specific taxa at ST2 were Calciosolenia brasiliensis, Nitzschia sp., Ophiaster sp., and nano-coccolithophores (ESC 10–20 µm). The highest number of specific species was found at ST3, and most of them were diatoms: Chaetoceros contortus, Ch. convolutus, Ch. debilis, Lennoxia faveolata, N. sicula, Proboscia alata, Pseudo-nitzschia pseudodelicatissima, R. hebetata f. semispina, R. cleveii, Thalassionema nitzschioides, and nano-scale Thalassiosira sp. (ESC <20 µm). Other two specific taxa for ST3 were Micromonas sp. and Oxytoxum cf. variabile (ESC <20 µm) (Table 3). Highest abundance of diatoms at ST3 was recorded thanks to the high quantities of Pseudo-nitzschia pseudodelicatissima.

Correlation between pigments and phytoplankton counts

HPLC based pigment concentrations closely followed the significant across-transect trends observed in phytoplankton abundances, as demonstrated by the RELATE test (Fig. S1B). Alloxanthin, zeaxanthin, DVChl b, and lutein contributed the most to similarities in trends, as shown by the BEST global test (Table S3). A clear split between coastal, eutrophic ST3 and oligotrophic ST1 and ST2 is visible in dendogram visualising LINKTREE constrained binary divisive clustering analysis done on phytoplankton counts and pigment concentrations (Fig. 5). This primary split (Node A>B, K), that can be explained by the specific threshold of alloxanthin (−0.0089 µgL⁻¹ for ST1 and ST2, 0,001 µgL⁻¹ for ST3) is highly significant (SIMPROF test, Fig. S1A).

Further splits in the dendogram are driven by secondary pigments and demonstrate finer differences within the ecosystem types, on oligotrophic side lutein or zeaxantin (Node B), and on eutrophic side by zeaxantin (Node K>L, N) and further down (Node N) by the lutein. Note that only some of the splits in this dendogram are significant (black lines on the Fig. 5) according to SIMPROF test (Fig. S1A).

Horizontal distribution of phytoplankton

Planktonic dispersal rate varies across marine planktonic taxa, while negative relationship between dispersal scale and body size causes less abundant and larger-fraction plankton (in near-surface, epipelagic waters) to have shorter dispersal scales and larger spatial species-turnover rates than the more abundant, smaller-fraction plankton (Villarino et al., 2018). The larger phytoplankton will be more similar at geographically proximate locations, and dissimilar between distant locations while it would allow smaller, more abundant phytoplankton (body size <2 mm) to travel greater distances (Finlay, 2002; Martiny et al., 2006; Villarino et al., 2018). This explains even spatial distribution of all nanophytoplankton fractions between stations, while microphytoplankton fractions, especially diatoms, are most abundant at ST3 (Figs. 3A and 3B). Additionally, we observed the highest number of specific diatom species at ST3 (Table 2).

Phytoplankton community structure

Phytoplankton community of North Pacific was comprised of 207 taxa, of which most abundant were coccolithophores (35.5%), then diatoms (25.2%), dinoflagellates (19.5%) and others (19.8%) including cryptophytes, phytoflagellates, and silicoflagellates. Microphytoplankton was more abundant at eutrophic ST3, with diatoms being dominant taxa while oligotrophic ST2 and ST1 were dominated by nano-scale coccolithophores and dinoflagellates, which is supported by molecular data obtained from the same cruise shown in Durkin et al. (2022). Proportion of taxa read abundances in exported ASVs (Amplicon sequence variants) was distinct between ST2 and ST3, with higher diatom read abundances at ST3 (78%), and lower at ST2 (10%), and small proportion of cryptophyta and haptophyta reads at ST3 (15%), and large contribution at ST2 (49%) (Durkin et al., 2022).

Microphytoplankton

Microphytoplankton abundance increased at eutrophic ST3 (Fig. 2A), where diatoms were dominant (Table 3) and contributed the most to the dissimilarity to other stations (Table 2). Similar assemblage was discussed in a study done by Iriarte & Fryxell (1995) who researched microphytoplankton community structure at equatorial Pacific at 140°W during El Niño 1992 event. Taxa groups that contributed the most to the biomass were diatoms, dinoflagellates and coccolitophores. Dominant species during March to April were Pseudo-nitzschia delicatissima, Thalassionema spp., Thalassiothrix spp., Thalassiosira lineata, and Oxytoxum variabile. In October the same species dominated, with additional two: Calcidiscus leptoporus and Chaetoceros atlanticus. Furthermore, Yamaguchi et al. (2002) analyzed plankton of three regions in western North Pacific: subarctic, subtropical and transitional. Eutrophic subarctic region had the highest phytoplankton biomass, and community dominated with dinoflagellates and diatoms. During North Pacific cruise along 155°W in January 1966, diversity of diatoms was analyzed using light microscopy, and 54 species were identified, of which 37 were frequent (Venrick, 1971). Dominant diatoms were Chaetoceros atlanticus, Ch. peruvianus, Denticula semina e, Nitzschia lineola, Thalassiothrix longissima, and Pleurosigma normanii (Venrick, 1971). Moreover, analyses of time series database including 12-year record of phytoplankton abundances, yielded list of dominant species comparable to ours: N. bicapitata , N. braarudii, N. sicula, Pseudo-nitzschia pseudodelicatissima., Thalassionema sp,.and Oxytoxum cf. variabile (Venrick, 1990). A recent molecular study on diatom assemblages in CCS revealed species with highest relative abundance of reads: Rhizosolenia sp., Actinocyclus sp., Thalassiosira diporocyclus, Asteromphalus sp., and Fragilariopsis doliolus (Abdala et al., 2022).

Eutrophic ST3 had the highest abundances of Pseudo-nitzschia pseudodelicatissima that was absent from ST1 and ST2. Pseudonitzschia taxa, while cosmopolitant (Hasle, 2002), seems to be prevalent in communities along the California coast, where it has been recorded since 1930s (Gran & Thompson, 1930). The appearance of these species coincided with upwelling zones near the coast (Trainer et al., 1998), while others point to areas with increased fertilizer use and agricultural run-off causing eutrophication (Smith et al., 1990). Pseudo-nitzschia species are known to respond to the environmental drivers—both human induced and those specific for ecosystem (Parsons & Dortch, 2002), and increased molecular and taxonomy analyses yielded more knowledge on their ecology, physiology, phylogeny and distribution (Trainer et al., 2012). Species recorded in CCS often causing toxic blooms are P. australis and P.multiseries, while P. pseudodelicatissima was frequent along Washington coast (Trainer et al., 2012). A diatom species Lennoxia faveolata had the second-highest abundance among diatoms at ST3 and was not detected in other stations. (Thomsen et al., 1993) who first described it, found high numbers in samples from Californian waters during winter, but not much more is known about it.

Coccolithophores

Coccolithophorid contribution to community composition is significant on all stations, with dominant nano- fraction, especially at greater depths. Micro-scale coccolithophores have a more significant abundance at ST1 and ST2, but they are absent at ST3. Coccolithophorid pigment proxy 19HF has a relatively high ratio on all stations when compared to other pigments. Its presence may point to the higher contribution of pico-scale coccolithophores in bigger depths at ST3. Michaelsarsia adriaticus was a dominant species present only at ST1, whereas dominant Calciosolenia brasiliensis and Ophiaster sp. were specific to ST2. Dominant species observed on both ST1 and ST2 were C. murrayi and Discosphaera tubifera. ST3 was dominated by nano-scale coccolithophorids (ESC <5 µm and 5–10).

Domination of coccolithophores at ST1 and ST2 point to species more adapted to oligotrophic conditions, while indirect observation of 19HF at ST3 implies a shift to the more eutrophic-adapted, smaller coccolithophores species. Li et al. (2013) observed concentrations of 19HF in the Pacific that was low in the upper euphotic zone but increased with depth. Fujiki et al. (2016) also observed low surface 19HF and 19BF concentration (<0.5 mg m⁻³) that is increasing below 70 m. This would suggest that the coccolithophores are physiologically adapted to low light, nutrient-enriched regions of the water or the 19HF came from other lineages containing the coccolithophorid-indicative marker pigment (Carreto et al., 2001; Landry, 2003).

Okada & Honjo (1973) recorded 90 coccolithophorid species in North and Central Pacific. Based on community structure, they described six zones, and Zones B (Transitional Pacific) and C (Central-North Pacific) match our sampling transect. They observed high abundance of Emiliania huxley (cold variety), Rhabdosphaera clavigera, and Umbellosphaera irregularis. Less abundant species present in this area were Discosphaera tubifera, Syracosphaera spp, R. stylifera, U. tenuis, Umbilicosphaera hulburtiana, U. sibogae, which is similar community structure observed in our study (Table S1), however we observed other dominant species: Calciosolenia brasiliensis, C. murrayi, D. tubifera, Michaelsarsia adriaticus, and Ophiaster sp (Table 3). Hoepffner & Haas (1990) identified nanophytoplankton community of NPSG by using electron microscopy, and observed that Prymnesophyceae contributed the most (55%), with equal abundances of Prymnesiales and Coccosphaerales. Dominant taxa were E.huxley, O. formosu s, R. clavigera, and C. murrayi. Another study in NPSG revealed a total of 53 species, from which most abundant were D. tubifera, U. tenuis, and Heladosphaera cornifera in the upper layer (0–80 m), and Anthosphaera oryza, Florisphaera profunda, Thorosphaera flabellata, and Oolithotus fragilis in the lower layers (140–200 m) (Reid, 1980). Other frequent species observed at the surface were Acanthoica acanthifera, Calyptrosphaera oblonga, U. irregularis, R. stylifera, S. pulchra, and S. pirus (Reid, 1980). Most of listed species we detected as well (Table 3).

Phytoplankton chemotaxonomy and its relation to microscopy

Chemotaxonomy is a method that allows characterization of the phytoplankton community to coarser taxa than the microscopy can, however, offering insight into the nano- and pico-planktonic composition that is undetectable by classical microscopy methods (Kramer, Siegel & Graff, 2020). Following the decades of research in which pigment composition was related to the microscopy based one, it proved to have biases as concentrations of pigment biomarkers, and their relation to the chlorophyll a are not always the best representative of the targeted taxa (Havskum et al., 2004; Irigoien et al., 2004; Pan et al., 2020). These vary with physiology of the cells, and environmental factors such as irradiance, nutrient availability, day length, temperature, and mixing status (Higgins, Wright & Schluter, 2011).

Regardless of biases, HPLC approach allowed us to track distribution of cyanobacterial taxa indirectly through their pigments proxy, DVChl a, and zeaxanthin, that reached their highest peak at oligotrophic ST1 and ST2. Highest concentrations of fucoxanthin, peridinin, 19HF, alloxanthin, and prasinoaxanthin were observed at ST3 (Table S4), which indicate higher abundances of diatoms, dinoflagellates, coccolitophores, cryptophytes, and prasinophytes respectively. Similarly, chemotaxonomy analyses of subsurface chlorophyll maximum (SCM) at western Baja California also showed dominance of diatoms, prasinophytes, and cryptophytes, in addition to haptophytes, pelagophytes, and picocyanobacteria (Almazán-Becerril, Rivas & García-Mendoza, 2012). Recent HPLC 22-year time series data on CCS phytoplankton community confirms significant contribution of diatoms, followed by dinoflagellates, chlorophytes, prymnesiophytes, and picophytoplankton (Catlett et al., 2021). Furthermore, one interesting trend arose from LINKTREE statistical analysis, showing that pigment alloxanthin determined most the differences in phytoplankton community between the CCS and open ocean stations (Fig. 5). While the imaging-based analysis did point cryptophytes play an important role in distinguishing two communities (Table 2), other taxa, namely coccolithophorids and diatoms also seemed to drive the ecosystem differences.

Claustre (1994) proposed another use of pigments that can determine trophic state of the area by calculating the F_p index (the ratio of pigments highly correlated to changes in Chl a concentration to other pigments that are less variable). Study showed fucoxanthin and peridinin, biomarkers for mostly diatoms and dinoflagellates respectively, had a higher correlation with the change of total Chl a in comparison to other pigments, meaning rise in biomass can be correlated with diatoms and dinoflagellates growth. We used the same approach and calculated higher F_p index at ST3 in comparison to ST1 and ST2, which correlates with the rise of fucoxanthin and peridinin in CCS (Table S4).

Contrasting North Pacific ecosystems

In this paper, the analyzed data showed distinct environments characterized by differences between phytoplankton abundances and concentrations of pigments along a transect that comprises an open ocean and a coast. We recorded lower phytoplankton counts at NPSG which is the largest ecosystem on the planet with reduced intake of nutrients in the euphotic zone, low primary production (PP) and export of carbon to deeper layers (Karl & Church, 2017; Kavanaugh et al., 2018). On the other hand, the phytoplankton counts were higher at the CCS that is more eutrophic ecosystem with seasonal upwellings, higher PP, and frequent blooms (Monterey Bay Aquarium Research Institute, 2022; Closset et al., 2021). Furthermore, the Columbia River influences the CCS with an increased terrigenous contribution raising the trophic state (Hickey & Banas, 2008). Differences between ecosystems can also be detected by observing maximum chlorophyll a concentration, and our results show that it was higher and more variable in CCS, in comparison to NPSG. High chlorophyll a concentrations were already recorded in northern part of CCS (Ware & Thomson, 2005), while studies in other eutrophic ecosystems show similar trend (Zhang, Wang & Yin, 2018; Miranda-Alvarez et al., 2020). Fujiki et al. (2016) observed low Chl a concentration at the ALOHA station (<0.05 mg/m³), whereas at ST1 and ST2 concentration did not exceed 0.4 µg/L.