Picophytoplankton size and biomass around equatorial eastern Indian Ocean

Abstract The cellular size and biomass of picophytoplankton were studied by flow cytometer during spring monsoon (March–May of 2015) in equatorial eastern Indian Ocean. We established an empirical relationship between forward scatter and cellular size to address the size and biomass of picophytoplankton. Results indicated that mean cell diameter of Prochlorococcus (0.60 μm) was the smallest, and then followed by Synechococcus (0.98 μm) and picoeukaryotic phytoplankton (1.05 μm). Thereafter, the biomass converted by abundance reached 0.64 μg·C·L−1 for Prochlorococcus, 0.34 μg·C·L−1 for Synechococcus, and 0.20 μg·C·L−1 for picoeukaryotic phytoplankton. Additionally, the distinct biomass contribution of picophytoplankton appeared to be affected by abundance, but not changes in cellular size. Vertically, the cellular sizes of picophytoplankton were remarkably small in upper waters, which was predominantly controlled by the nutrient availability. In contrast, they were larger in deeper waters, which was primarily attributed to the combined effects of low temperature and reduced light availability. Spatially, under the influence of high nutrient concentration induced by the different circulations and coastal upwelling, slightly high carbon biomass of picophytoplankton was observed around the coastal zones of Sri Lanka island and Sumatra, as well as the southern Bay of Bengal.

terms of the size and biomass of picophytoplankton. Thus, presenting their size and biomass is critical to understand the contributions to carbon cycles of these special taxa in the Indian Ocean.
Thus far, flow cytometer (FCM) can help us to address the size and biomass of picophytoplankton at high frequency according to their cell morphological properties and fluorescence when the highsensitive protocol was used. For FCM, light scattering at different angles are related to the function of particle volume and secondarily shape (Latimer, 1982). However, an empirical calibration between cell diameter and side scatter (SSC) was performed to roughly estimate equivalent spherical diameter and cellular biovolume of picophytoplankton (Calvo-Díaz & Morán, 2006;Chen et al., 2011).
The range of picophytoplankton cell diameter, in general, used to establish the empirical relationship between cellular size and SSC is still critical (Gasol & Del Giorgio, 2000). Light scattering efficiency of picophytoplankton cell is a complex function of its size, structure, and refractive index, even different FCM and fixatives may yield significantly different scatter diagrams of the same sample as a function of relatively minor changes in detection geometry (Gasol & Del Giorgio, 2000). Consequently, Allman, Hann, Manchee, and Lloyd (1992) pointed out that cell diameter and light scattering should break down when comparing different species. According to the Mie theory, when particle diameter extends from 0.2 μm up to 2-3 μm or more, forward scatter (FSC) is the signal which is the most sensitive to cellular size, with a diameter d dependence of FSC in d 4d 6 (Morel, 1991). For example, the converting mean FSC to cell sizes for Synechococcus was done by fitting a power relationship with laboratory calibrations (FSC = a × Diameter b ), and the exponent (b) was found to be 5.4, which was reasonably close to that determined by Mie light scattering theory (DuRand, Olson, & Chisholm, 2001); FSC versus biovolume data of Synechococcus reported a value of d 5.1 (Chisholm 1992); an FSC variation for Prochlorococcus during the daytime as a doubling in the average volume of the prokaryotes indicated a value of d 5.4 (Binder, Chisholm, Olson, Frankel, & Worden, 1996); the FSC related to particle sizes of reference beads for picoplankton led to the value around d 5 (Blanchot, André, Navarette, Neveux, & Radenac, 2001). Collectively, a strong correlation between FSC and cellular size has been determined by laboratory cultures of reasonably spherically shaped cells (DuRand et al., 2001;Olson, Zettler, & Anderson, 1989), despite small changes in refractive index (DuRand & Olson, 1998). Moreover, Koch, Robertson, and Button (1996) presented that FSC is chosen over SSC because of its far greater signal intensity to subcellular structure according to the theoretical basis of their approach. Actually, the relationship between FSC and picoplanktonic size has been carried out for several decades, even always involving bacterioplankton. Typically, Robertson and Button (1989) have made use of FSC to estimate bacterial size and proposed a good relationship between FSC and bacterial volume. However, some recently published papers often showed that the same large dispersion of beads and target cells in FSC limited the application of FSC and weakened its relationship with cellular size. To capture light scatter in forward angles and increase the sensitivity of this parameter, more instruments have been equipped with photomultiplier tubes (Gasol & Del Giorgio, 2000). This led Blanchot et al. (2001) to attempt a practical way for the estimation of mean cellular size what the relationship between the mean FSC and cellular diameter was determined by the power law empirically, which was assumed to stand for the mean FSC and diameters relative to those of the beads (d cell = d bead (FSC) 1/5 ), respectively.
In this study, we established an empirical relationship between FSC and cellular size to address the cellular size and biomass of picophytoplankton, and then to understand more clearly whether and how the contrasting environmental conditions affect their variations in equatorial eastern Indian Ocean. More specifically, will the expected environmental conditions induced by the complex circulations and water masses change the cellular size and biomass of picophytoplankton? Plus) rosette sampler. Photosynthetically active radiation (PAR) was measured by an RBR sensor (XRX-620). The euphotic depth was defined as the depth of 1% surface light penetration. Temperature and salinity were recorded at the same time.

| Sampling strategy
Seawater samples for picophytoplankton analysis by FCM were preserved on board with paraformaldehyde (1% final concentration).
To avoid loss of resolution and changes in cell size due to fixation or freezing, FCM samples were kept in the dark without treatment at room temperature for 10-15 min, and then quickly freeze-trapped in liquid nitrogen until analysis in the laboratory (Den Engh et al., 2017;Sommaruga, Hofer, Alonso-Sáez, & Gasol, 2005).
Samples for nutrient analysis were filtered using 0.45μm cellulose acetate membrane filters, and then immediately refrigerated at −20°C for further analysis. Nutrient concentrations including ammonium, phosphate, nitrate, nitrite, and silicate were performed by a Technicon AA3 Auto-Analyzer (Bran + Luebbe) according to the classical colorimetric methods. Dissolved inorganic nitrogen (DIN) defined as ammonium+nitrite+nitrate was analyzed using the copper-cadmium column reduction method. Dissolved inorganic phosphorus (DIP) and silicate (DSI) were measured using the typical spectrophotometric methods (Dai et al., 2008;Guo et al., 2014).
In all ranges of tested low nutrient standards, the AA3 was more precise and more accurate and showed lower detection limit for all channels: 0.018 μmol·L −1 for DIP, 0.009 μmol·L −1 for nitrate + nitrite, and 0.012 μmol·L −1 for DSI (Dafner, 2015).

| Flow cytometry analysis
Within the present study, three dominating populations, namely Synechococcus, Prochlorococcus, and picoeukaryotic phytoplankton, were manually distinguished by FCM (BD Accuri C6) according to their different amplitudes, shapes, and optical signals. The aforementioned references revealed that FSC was much more adaptable to establish the empirical relationship with cellular sizes of picophytoplankton (<2 μm). Thereafter, cellular diameters of them were enumerated by using the FCM based on their distinct FSC signatures. The histograms of example FSC frequency of three picophytoplankton groups are shown in Figure 2. We estimated the sizes of mean FSC frequency were of the order of the cell diameters of Prochlorococcus, Synechococcus, and picoeukaryotic phytoplankton, respectively. Otherwise, we presumably assumed that the optical signals of mean cell diameters of Prochlorococcus (0.6 μm) were similar to the normalized beads; subsequently, the cellular sizes with respect to biovolumes were estimated on the basis of the mean FSC frequency of Prochlorococcus, Synechococcus, and picoeukaryotic phytoplankton as measured relative to those of the mean cell diameters (Partensky, Hess, & Vaulot, 1999). The empirical relationship between the mean FSC (MFSC) and cell diameters (d cell ) is shown as follows: d cell = d bead(Prochlorococcus) (MFSC cell /MFSC Prochlorococcus ). This hypothesis that was similar to the pattern of Blanchot et al. (2001) was an oversimplification, although the result needed to be considered cautiously and more proper calibration in further studies, the relationship was in accordance with those presently quoted.
Additionally, two different subclusters of Synechococcus were directly evidenced by two distinct peaks (bimodal distribution) in the histograms of the FSC frequency (Figure 2b), and possibly corresponding to the "SynechococcusI" and "SynechococcusII" cells (Zhao et al., 2013). Unfortunately, due to the absence of distinctive features and peak-overlap, it was difficult to identify them by the FSC frequency. Therefore, these subclusters in the picophytoplanktonic fraction of the Indian Ocean were artificially combined and exclusively represented by Synechococcus.

| Cell size and biomass
Results from the estimated mean FSC method indicated that mean cell diameter of Prochlorococcus (0.60 ± 0.22 μm) was the smallest, followed by Synechococcus (0.98 ± 0.44 μm) and picoeukaryotic phytoplankton (1.05 ± 0.30 μm). Our estimated cellular sizes for picophytoplankton were comparable with other estimates.  represented vertical variations of cell size of three picophytoplankton groups, respectively. In particular, the observed vertical trend of cell size of picoeukaryotic phytoplankton was very similar to the pattern in the Arabian Sea (Shalapyonok et al., 2001), which showed that picoeukaryotic phytoplankton in the surface were slightly larger than those in the subsurface chlorophyll maximum. Diameter Picophytoplankton biomass could be measured directly from proximate analyses (such as chlorophyll, ATP, carbon, and nitrogen concentration) or indirectly from biovolume characteristics of the enumerated population (Frame & Hu, 1990;Hewes, Sakshaug, Reid, & olm-Hansen, 1990;Hunter & Laws, 1981;Sun & Liu, 2003).
Nevertheless, to separate Prochlorococcus, Synechococcus, and picoeukaryotic phytoplankton from other algae cells, and then to estimate the different biomass of them, respectively, using proximate analyses in situ were very difficult because of their considerably small cells. Determinations of proximate constituents such as levels of chlorophyll or detritus plus viable carbon were, therefore, inadequate measurements for use in studies on species or size-class contributions to primary production (Sun, Liu, & Qian, 2000). For each picophytoplankton assemblage, the mean biovolume should be calculated from the mean value of these individual cell biovolumes, rather than directly from the mean cell diameter. In the equatorial EIO, the average abundances of Prochlorococcus, Synechococcus and picoeukaryotic phytoplankton were at the magnitude of 10 3 -10 5 cells·ml −1 (Table 1); hence, it was a extremely complicated and heavy work to calculate the biovolume of each picophytoplanktonic cell. However, mean cell size differs greatly between species of picophytoplankton that suggests the mean cell size of species is very important with biomass (Agusti, Duarte, & Kalff, 1987;Liu, Chang, Tseng, Wen, & Liu, 2007;Marañón, 2015). Accordingly, Hillebrand, Dürselen, Kirschtel, Pollingher, and Zohary(1999) proposed that the biovolume can be calculated from the mean of measured cell diameters, not as a mean of a set of individually calculated biovolumes although there may be some errors. Indeed, the error sources within this study primarily came from the choice of mean cell diameter, in addition to the accuracy of measurement and consequent estimation of biovolume. When the two methods for mean biovolume calculation were compared, Sun and Liu (2003) found that although the latter method usually underestimated the variability, its trend had better agreement with increased measurements. Under most circumstances, the standard error was <5% of the mean biovolume after the measurement of 10 cells, we suggested that taking as many measurements as possible was better. As yet, the conversion of carbon content from biovolume based on mean cell diameter is the good way to estimate the gross biomass of picophytoplankton in the absence of direct unicellular size measurement. For example, Calvo-Díaz and Morán (2006) obtained the mean diameters of the different groups of picophytoplankton by an empirical calibration, and thereafter assumed a spherical shape for all groups to estimate biomass by using the method of biovolume-to-carbon conversion in the southern Bay of Biscay; Chiang, Kuo, Chang, Wang, and Gong(2002) computed the mean cell volume based on the approximately coccoid shape of Synechococcus cells to estimate biomass in the East China Sea; Blanchot et al.(2001) attempted a very practical way for the estimation of mean cell size to calculate cellular carbon in the equatorial Pacific. Thus, the mean measured cell diameter of picophytoplankton could be used to calculate biovolume in routine analysis.
The carbon biomass estimates of Synechococcus, Prochlorococcus, and picoeukaryotic phytoplankton for this study area were roughly calculated by using the mean carbon content per cell multiplied by cell abundance for each of these groups. Preferentially, we assigned the geometric shape for a picophytoplanktonic cell as a sphere (Chisholm et al., 1988). The cell diameter (d) was converted to biovolume (V) using a predictive equation (Sun et al., 2000): . with those reported in the regions of tropical and subtropical Pacific Ocean, Prochlorococcus was the most abundant photosynthetic organism, even accounting for 65% of the total picoplankton biomass, whereas Synechococcus and picoeukaryotic phytoplankton constituted less than 35% of the biomass (Blanchot & Rodier, 1996;Campbell, Nolla, & Vaulot, 1994;Charpy & Blanchot, 1998).
Interestingly, some previous studies reported that Synechococcus could fix an order of magnitude more carbon than Prochlorococcus cells due to Synechococcus was slightly larger than Prochlorococcus, indicating that the relative biomass contributions of three picophytoplankton groups were not simply determined by abundance but also changes in cellular size (Flombaum et al., 2013). However, our finding demonstrated that the distinct biomass contributions of picophytoplankton in the equatorial EIO appeared to be affected by abundance, but not changes in cellular size. In addition to the south-

| Factors controlling on cell size and biomass
As reported previously for picophytoplankton, the differences in chlorophyll fluorescence signals (cellular sizes) observed with depth in water column were most likely due to photoacclimation (i.e., reduction of the pigment content at high light levels), cellular division, changes in quantum yield, and shifts in species composition (Campbell & Vaulot, 1993). Specifically, marine environmental variables such as temperature, light, and nutrient availability usually had apparent effects on the cellular sizes of picophytoplankton (Chen et al., 2011). Taken together, the cellular size of picophytoplankton appears to be under complex physiological and environmental control. In equatorial EIO, there were less nutrients in upper waters (5-50 m) during spring monsoon (Table 2).   and Synechococcus in the deep euphotic layers were significantly larger than those in the upper euphotic layers (Liu et al., 2007), and these differences in cell size were attributed to the growth rate at different nutrient concentrations. In dilution experiments, Liu et al. (1998) reported that the growth rate of Synechococcus was much higher than Prochlorococcus with nutrient availability.
Furthermore, Shalapyonok et al. (2001) found that the cyanobacteria cells inhabiting the top layer were smaller than those inhabiting the deeper layer, which might be related to nutrient depletion, as well as photoacclimation. Actually, this hypothesis that lower nutrient concentration in upper waters induced smaller size was typically suitable for picophytoplankton in equatorial EIO  TA B L E 3 Spearman's rank correlation coefficients between environmental factors and carbon biomass of three picophytoplankton groups and minimal diffusion boundary layer thickness (Raven, 1998).
Collectively, the variations of cellular sizes in upper waters were closely related to the nutrient availability.
Due to phytoplankton that grow deeper in the water column need more pigment per cell to compensate for the decreased light levels, their cellular sizes are broadly larger under lower light intensity (Goericke & Welschmeyer, 1998  Furthermore, the refraction of light is largely dependent on the particles, suspended sediments, and also color dissolved organic matter. Accordingly, it was impossible to establish the linear correlation between cell size and light level directly owing to the variability of light irradiance. According to the nonlinear fitting curve between depth and light irradiance (Figure 7), the light intensity decreased with the increasing depth in the equatorial EIO (R 2 = .65252), indicating that the decrease in light irradiance could be effectively represented by the increasing depth. Consequently, we established the linear response between cell size and the increasing depth (representing the decreasing light irradiance) to show different response relations of cell size and light intensity. Figure 8 shows the distinct response relations between cellular size and depth (light levels) and temperature for three picophytoplankton groups in the equatorial EIO. According to our analysis of response, the slopes of these lines for cell size of Prochlorococcus and Synechococcus to depth and temperature were relatively high, which represented high response levels of cell size to light levels and temperature. Thus, this observation indicated the vertical patterns of Prochlorococcus and Synechococcus in cellular sizes were the results of combined effects of photoacclimation and temperature ( Figure 8). Similarly, Goericke and Welschmeyer (1998)  F I G U R E 8 Different response relations between cellular size and light levels (depth (m)) and temperature (°C) for three picophytoplankton groups were consequently critical in explaining the vertical variations of their cellular sizes in addition to temperature and light availability in deeper waters. For picoeukaryotic phytoplankton, however, due to the lower capacity to photoacclimate controlled by physiological differences, their cellular sizes were larger occurring in low light layers (Campbell et al., 1997). Similarly, the pronounced increase with depth in picoeukaryotic phytoplankton size between the depth of 100 m and 200 m was the result of a dramatic change in light intensity.
Spatially, most of coastal waters around the Sri Lanka island are influenced by the increases of pollution and eutrophication.
Simultaneously, the East India Coastal Current (EICC) along the western boundary of the Bay of Bengal flows equatorward and bifurcates east of the Sri Lanka island, but one bifurcation of its source waters characterized by nutrient enrichment continues along the coast of Sri Lanka island (Vinayachandran et al., 2005). Therefore, this studied coastal area is abundant in nutrients and suitable for picophytoplankton development. The CCA analysis revealed that significant correlations were found between the average carbon biomass of Prochlorococcus, Synechococcus, and picoeukaryotic phytoplankton and nutrient variables, reinforcing the notion that nutrient supply was the key limiting factor for the spatial distribution of average carbon biomass. Therefore, such contaminants loading from coastal currents and freshwater discharging from the EICC with high nutrients contributed to the fairly high average carbon biomass of three picophytoplankton groups in coastal waters of the Sri Lanka island. Surface temperature surrounding the Sumatra was below 29°C with a comparatively high salinity of approximately 34 during spring 2015, indicating that its hydrographic properties were prominently attributed to the coastal upwelling. The co-occurrence of high average carbon biomass for picophytoplankton in coastal upwelling zones nearby the Sumatra was not surprisingly owing to the high nutrient availability.
Waters of very low salinity and temperature (32°C and 29°C, respectively) were presented in the northern Bay of Bengal, thereby indicating that they were dramatically influenced by freshwater from the Bay of Bengal runoffs. As such, slightly high average carbon biomass of three picophytoplankton groups co-occurred in the southern Bay of Bengal due to they could benefit from the increased nutrient availability carried by the Bay of Bengal runoffs (Mukhopadhyay, Biswas, De, & Jana, 2006). Collectively, as a result of differential responses of picophytoplankton in average carbon biomass to the changes in physical and chemical environments induced by the variable circulations and water masses, we observed very different spatial patterns of average carbon biomass for these three picophytoplankton groups.

ACK N OWLED G M ENTS
The authors are grateful to D. Wang for supplying the temperature and salinity data for the cruise. We also express thanks to the crews of R/V Shiyan I. This study was supported by the National Natural Science Foundation of China (41676112 and 41276124); the National Foundation of the Indian Ocean Opening Voyage Project (41449910), and the Changjiang Scholar Program of Chinese Ministry of Education to J. Sun.

CO N FLI C T O F I NTE R E S T
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.