On-board flow cytometric determinations of picoplankton abundance (i.e. Synechococcus spp., Prochlorococcus spp., picoeukaryotes and also heterotrophic bacteria) were obtained in the East China Sea in fall of 2000 and 2003. The average abundances of Synechococcus, Prochlorococcus, picoeukaryotes and heterotrophic bacteria were 105, 105, 104 and 106 cells ml−1, respectively. Synechococcus, picoeukaryotes and heterotrophic bacteria were abundant at all the stations and presented higher concentration in the inner shelf where influences from the Changjiang effluent plumes and the coastal upwelling were evident, while Prochlorococcus was absent from the near-shore stations and became the dominant picophytoplankton population in offshore waters, where its abundance was comparable to that for heterotrophic bacteria. All picoplankton groups showed a reduction in cell number with depth, and a positive correlation with water temperature were observed, which reflected the importance of light and temperature on picoplankton growth. A negative relationship with salinity was found for heterotrophic bacteria along two sections across the East China Sea Shelf, and distribution of picoplankton was dominated by different water masses. The fixation could lead to loss in Prochlorococcus cell numbers within one month, and all the picoplankton numbers decreased dramatically after three months.
Continental shelves represent <10% of the surface area of the world ocean, but they are an important intermediate zone between the land and the open ocean, where terrigenous and anthropogenic materials are transported, deposited, and transformed. The continental shelf is characterized by important standing stocks of organic carbon and high rates of primary and secondary productions, and is recognized as one of the most important compartments of the global biogeochemical cycles of carbon and other biogenic materials .
The East China Sea is a marginal sea with one of the most extensive continental shelves in the world, interacting with large rivers (e.g. Changjiang) and the Kuroshio in its western and eastern boundaries, respectively. The Kuroshio interacts actively with shelf waters through frontal and upwelling processes (e.g. filaments and eddies). Four water masses can be commonly differentiated within the East China Sea Shelf: the relatively fresh, cool and nutrient-rich Changjiang Diluted Water, the oligotrophic and highly saline Taiwan Current Warm Water that reaches as far as the Changjiang (Yangtze River) Estuary, the intrusion of cold and nutrient-rich Kuroshio Subsurface Water, and the warm, saline and nutrient-poor Kuroshio Surface Water (Fig. 1). The major contribution to the nutrients in the East China Sea Shelf includes the Changjiang Diluted Water and Kuroshio Subsurface Water [2–4]. The nutrient supply via these pathways sustains high primary production in the East China Sea (e.g., 145 g C m−2 y−1) .
The operational concept of picoplankton (<2μm) includes by definition the autotrophic cyanobacteria Synechococcus spp.  and Prochlorococcus spp. , small eukaryotic algal groups, and heterotrophic bacteria , which are important components of marine plankton communities. Synechococcus is found ubiquitous in the upper temperate and warm ocean , and Prochlorococcus has been found to be more abundant in oligotrophic than in eutrophic waters [10,11]. These two phytoplankton groups, together with picoeukaryotes, have fast growth rates matched by high mortality losses caused by microzooplankton grazing, making them fundamental components of the biomass and primary production of marine ecosystems, and hence participating in nutrient regeneration and cycling in the ocean . The widely distributed heterotrophic bacteria and their variability can parallel that of phytoplankton within ecosystems [8,13]. Heterotrophic activity contributes to the conversion of dissolved organic matter (DOM) to bacterial biomass that can be transferred to higher trophic levels through the marine microbial food web with participation of flagellate and ciliate grazers . The quantification of picoplanktonic organisms is therefore of great importance for the characterization of marine ecosystems and for understanding the function of marine food webs.
Picoplankton groups in the East China Sea have been the subject of research in the past [3,15–25]. However, few data have been published on the distribution of picophytoplankton and bacterioplankton in the shelf. Previous studies were mainly focused on only one picoplankton group such as Synechococcus or on heterotrophic bacteria rather than simultaneously observing the four picoplankton groups, and most of the data in literature were obtained by epifluorescence microscopy, which was not as efficient, sensitive and precise as flow cytometry (FCM)  except in the study by Jiao et al. . It seems that none of the analyses with flow cytometry was made on board.
On-board simultaneous determination of these picoplankton groups was made during the fall of 2000 and 2003. The aim of this study was to better understand the relationship between the spatial structure of hydrographic properties of different water masses and the variability of different picoplankton groups, and to discern the factors controlling the distribution of picoplankton in different parts of shelf ecosystems.
2Materials and methods
2.1Station locations and sampling
Two cruises were conducted; cruise 2000 from October 20 to November 8, 2000, and cruise 2003 from September 4 to 26, 2003, on board R/V “Dong Fang Hong 2”. Cruise 2000 had five stations located along the PN section (named by Nagasaki Marine Observatory). On cruise 2003, the stations were set on three different but representative sections: PN section across the East China Sea shelf, AS section along the Okinawa Trough and YT section from the Changjiang Estuary to the Tsushima Strait, and were sampled twice within 20 days (Fig. 1). Stations P4 and A6 were not sampled in the first leg, and station P3 was not sampled in the second leg because of a typhoon. In 2000, Synechococcus spp., Prochlorococcus spp., and picoeukaryotes were enumerated, and in 2003, heterotrophic bacteria were counted as well. Temperature, salinity, turbidity, epifluorescence and dissolved oxygen were recorded in the water column using a Sea-Bird 911 plus conductivity-temperature-depth (CTD)-Rosette assembly. Seawater was sampled with Niskin bottles at different water depths designed from the CTD profiles.
2.2Analysis of picophytoplankton and heterotrophic bacteria
Since fixation with added chemical reagents may result in loss of cells , samples were analyzed on a FACScan flow cytometer (Becton Dickinson, San Jose, CA, USA) equipped with an air-cooled argon laser (488 nm, 15 mW), placed on-board so that after sample collection the analyses could immediately be performed. Forward light scatter (FSC), side light scatter (SSC), green fluorescence (530 ± 15 nm, FL1), orange fluorescence (585 ± 21 nm, FL2) and red fluorescence (>650 nm, FL3) were recorded for each particle in the sample, and data obtained were processed with CELLQuest™ software (Becton Dickinson, San Jose, CA, USA). Yellowish green fluorescent beads (1.002 μm) (Polysciences Inc., catelogue # 18660) were added to calibrate cell fluorescence emissions and light scatter signals, which allows the comparison of fluorescence and cell sizes among samples. The picophytoplankton groups could be discriminated and enumerated according to their specific autofluorescence properties and light scatter differences . For the enumeration of heterotrophic bacteria, fresh water samples were stained with 1:10,000 (vol:vol) SYBR Green I (Molecular Probes, Inc.), and incubated in the dark at room temperature for 15 min before analysis . Bacteria were measured for their side light scatter and green fluorescence signals, which were related to cell size and nucleic acid content, respectively . Triplicate measurements were made for each sample with precision higher than 7.2% (relative standard deviation).
To test the effects of preservation, duplicate samples were fixed for 15 min with different fixatives including paraformaldehyde (final concentration: 1%), glutaraldehyde (final concentration: 0.125%) and a mixture of paraformaldehyde and glutaraldehyde (final concentration: 1% and 0.125%, respectively), then deep frozen in liquid nitrogen [26,28] and analyzed after one and three months, respectively. All the stock solutions except the dye were pre-filtered through a 0.2 μm pore size filter before use to avoid contamination.
Statistical analysis of data was done with the SPSS software (SPSS Inc.).
In general, the coastal region was affected by fresh water with low salinity discharged from the Changjiang in fall 2000. In the surface water (0–30 m), hydrographic parameters were vertically well mixed, with temperature increasing from 21 to 25 °C and salinity from 26.5 to 34.0, from off the Changjiang Estuary to the edge of the shelf. There was vertical increase in salinity and decrease in temperature with depth, which was more pronounced in the outer shelf waters where a pycnocline at a depth of about 40 m was formed. There was a salinity gradient from the shore to the mid-shelf areas (i.e., salinity increasing from 26 to 33), which is quite distinctive in summer when the Changjiang effluent plume spreads over the continental shelf. In the bottom waters, an area of uniform salinity (34.4–34.6) was formed but with a strong vertical temperature gradient (15–22 °C) in the outer shelf. This could be an evidence of the intrusion of the cool and saline Kuroshio Subsurface Water.
In 2003, the average surface water temperatures were >25 °C at all stations, and the surface salinity increased from the inner shelf (23.0) to outer shelf (34.4) in early fall. Both water temperature and salinity in the PN and YT sections showed similar distributions (Fig. 2(a)), which were comparable to that of 2000. However, an upwelling of cold and saline water could be seen just outside the Changjiang Estuary at a depth of about 25 m in 2003 (Fig. 2(a)). As a result of the mixture of the Changjiang Diluted Water and the Kuroshio with strong surface heating by radiation, a pycnocline developed in the PN section and became deeper towards the shelf break (Fig. 2(a)). There was cold upwelling water at station T5 as well (Fig. 2(a)), which might indicate the shelf mixing water. In the AS section, the water temperature decreased steadily with depth until to <5°C in bottom, while the salinity increased with depth (i.e. 33.3–34.9; Fig. 2(a)). There was a high salinity area in the subsurface water near station A6, which seemed to be the contribution of the Kuroshio Subsurface Water (Fig. 2(a)).
3.2Distribution of picoplankton in two cruises
In most of the samples, three groups of picophytoplankton could easily be discriminated in unstained samples according to their autofluorescence properties and light scatter differences. Also, at least two groups of heterotrophic bacteria, each with distinctive green fluorescences related to nucleic acid content, could be distinguished after DNA staining (Fig. 3). These results were in agreement with studies in other world areas [11,26,28,30].
On both cruises, the concentrations of Synechococcus ((59.7 ± 98.0) × 103 cells ml−1] and Prochlorococcus ((47.3 ± 101.8) × 103 cells ml−1] were comparable and were one order of magnitude higher than that of picoeukaryotes ((5.5 ± 7.7) × 103 cells ml−1] in the PN section. In the YT section, the concentration of Synechococcus ((190.2 ± 527.7) × 103 cells ml−1] was one to two orders of magnitude higher than those of Prochlorococcus ((39.1 ± 82.4) × 103 cells ml−1] and picoeukaryotes ((7.3 ± 14.2) × 103 cells ml−1]. Overall, the distributions of three autotrophs were similar along these sections, that is, the abundances of the picophytoplankton groups tended to increase in surface waters in mid-shelf and decrease with depth, and were rarely detected beneath the euphotic zone (Fig. 4). Radiation was not limiting phytoplankton growth in surface water offshore the turbid Changjiang effluent plume [5,15]. The upwelling of the nutrient-rich Kuroshio subsurface water, which can be found in the middle shelf of the East China Sea  (Fig. 2(a)), stimulated growth of these photosynthetic autotrophs at the mid-shelf along the PN section in both years (Fig. 4). The supply of inorganic nutrients from the Changjiang runoff in combination with and the coastal upwelling, induced the development of high picoeukaryote biomass at station P10 and high cell numbers of three autotrophs at station T9 in 2003 (Fig. 2(a) and Fig. 4). It should be noted that Prochlorococcus was almost absent near the Changjiang Estuary (e.g. station P12 in Fig. 4). The Prochlorococcus was abundant farther seawards than Synechococcus and picoeukaryotes in the PN section in 2000 (Fig. 4). Also, the high abundance site of Prochlorococcus was separated from those of Synechococcus and picoeukaryotes in the PN and YT sections in 2003 (Fig. 4). Along the AS section, high abundance sites of autotrophs appeared in high salinity area influenced by the Kuroshio Subsurface Water (Fig. 2(a) and Fig. 4), where the abundance of Prochlorococcus was almost one to two orders of magnitude higher than those of Synechococcus and picoeukaryotes (Fig. 4). This was similar to the other reported results [11,20]. Furthermore, the numbers of autotrophs remained relatively high and stable in surface mixing waters down to the pycnocline, and then decreased sharply with depth (Fig. 4).
Cell abundances of heterotrophic bacteria over the East China Sea Shelf were often one to two orders of magnitude higher than those of picophytoplankton. The vertical distribution of bacteria decreased with depth, similar to distributions of autotroph groups. The presence of bacterioplankton tended to diminish with offshore distance (Fig. 4). A high abundance site formed also at station T5, where water was well mixed because of the upwelling induced by the shelf mixing water (Fig. 2(a) and Fig. 4).
4.1Effects of fixation
Fixation was reported to lead to cell loss [26,27]. In this study, the difference between the different fixatives was statistically insignificant, so only the data of samples fixed with paraformaldehyde are shown (Fig. 5). The cell numbers and fluorescence parameters by FCM were roughly unchanged during one month after sampling, except in Prochlorococcus, shown by the variation of slopes of the simulation lines in Fig. 5. After three months, the cell numbers of all picoplankton groups dramatically declined and data scatter considerably, particularly for Prochlorococcus and picoeukaryotes (Fig. 5). If the samples are not fixed, the abundances will decrease more dramatically and data will scatter significantly. It can be concluded that samples should be analyzed on board or kept in liquid nitrogen before analysis (i.e. within one month) once back to the laboratory.
4.2Comparisons among different sections and cruises
Station P12 was always affected by the Changjiang diluted water, and stations in the mid-shelf at PN section were mainly dominated by the Taiwan current warm water  on both cruises (Fig. 2(b)). Nevertheless, the distributions of water temperature and salinity were quite different in the inner shelf between the two years. Stratification was more evident, temperature was higher in the inner shelf, and salinity was higher in the subsurface and bottom waters in the middle shelf in the year 2003 compared to 2000. This may explain the differences in the locations of high abundance phytoplankton and cell numbers between the two cruises (Fig. 4), as there was a comparatively stronger effect from the Changjiang Diluted Water in 2003 (Fig. 2(a)); the nutrients from the Changjiang [2,3], together with a more favorable temperature, promoted the growth of autotrophs. It is possible that limited number of stations on the cruise in 2000 may have missed high biomass areas of varying picophytoplankton populations. In 2003, the four picoplankton groups had similar locations of high abundances in two legs along the PN section, showing a relative stability of cell abundances along with hydrographic features (Fig. 4).
The hydrographic properties in the YT section in 2003 are comparable to those in the PN section: affected by the Changjiang Diluted Water in near-shore waters and by the Taiwan Current Warm Water in offshore waters (Fig. 2(b)), and showing a remarkable similarity between the two legs. The exception is a lower salinity (t-test, p < 0.01) above 10 m at stations in the near-shore areas (e.g. station P12 to T8) for the first leg, which could be seen in picophytoplankton abundances at stations T9 (Fig. 6). Similar high densities for Prochlorococcus and picoeukaryotes in the first leg at station T9 were not found in the second leg. The thermocline was enhanced in the outer shelf at the YT section (i.e. station T3, Fig. 6) of the second leg, with stronger stratification preventing the supply of nutrients to the surface waters . The implication includes the lower concentrations of picoplankton groups.
Overall, influences of the Changjiang Diluted Water at the PN and the YT sections in the year 2003 are comparable (p < 0.01) with respect to the salinity data. The temperature and salinity in the surface water at station T5 (Fig. 2(a)) of the YT section were lower, indicating upwelling that was probably induced by the exchange of shelf mixing water with the Yellow Sea. The presence of an upwelling that can facilitate vertical mixing and bring up nutrients and DOM for Synechococcus and heterotrophic bacteria, respectively, is a plausible explanation for the higher concentrations of those in the mid-shelf area of the YT section as compared to the PN section (Fig. 4).
At the AS section, the hydrographic conditions were more stable, and the stations were affected by the Kuroshio Surface and Subsurface Water (Fig. 2(b)). Cell numbers were rather low and stable, except for Prochlorococcus that was more abundant in oligotrophic waters [10,11]. Synechococcus and picoeukaryotes seem to have a higher demand for nutrients and hence flourished in the Changjiang Diluted Water and in the upwelling rather than in the AS section. Heterotrophic bacterial numbers were also lower in the AS section than in the other two sections (Fig. 4), probably because the favorable conditions such as biodegradation of organic matters promote heterotrophic activity [14,17,25], and those conditions originated presumably from the discharge from the Changjiang . The high-salinity Kuroshio Subsurface Water incursion is a major nutrient source for the East China Sea , sustaining picoplankton growth with high abundance at station A5 (Fig. 4). Apparently, the uplift of the Kuroshio Subsurface Water at station A5 was enhanced in the second leg, as was seen from the difference in salinity (Fig. 6), with rather different distributions of Prochlorococcus and other picoplankton groups. Prochlorococcus cell numbers decreased with depth in both legs, while the other groups showed high numbers at depth of 60 m rather than at the surface in the second leg (Fig. 6), indicating that the Kuroshio Subsurface Water had an impact on other picoplankton groups. This implies that Prochlorococcus and the other photosynthetic groups in oligotrophic Kuroshio waters have different demands for nutrients.
4.3Comparison of picoplankton abundance
To avoid introducing additional biases related to differences in biomass conversion factors [19,21,22], cell abundances are compared in this study, and the biomass data from other studies are converted to abundances with the relevant conversion factor for each study.
The concentrations of different picoplankton groups obtained in this study are roughly of the same order as in the other studies conducted in the East China Sea (Table 1). The abundance of heterotrophic bacteria in this study is lower than that reported in literature [3,19,24,25], which may be merely because of different study areas, since high supply of DOM required for bacterial growth was reported at stations concentrated in the inner shelf . However, it was also reported that Prochlorococcus could hardly be separated from bacterioplankton by means of epifluorescence microscopy, and thereby might be mistaken for the latter . Such kind of bias can be significant especially in oligotrophic waters, where Prochlorococcus can reach nearly 20–40% of the total prokaryotes [10,30,32]. Furthermore, cell loss caused by chemical fixation in other studies  could have induced artifacts in previously published data on picoplankton (cf. Fig. 5). Our on-board measurements using FCM provide more reliable information for the discrimination of picoplankton groups.
Table 1. Abundances of picoplankton groupsa in the East China Sea from different studies
Cell numbers (average ± standard deviation; 103 cells ml−1)
4.4Relationship between picoplankton and hydrographic properties
Although radiation effects are weak in the fall , there was a dramatic impact of temperature on cell numbers of all picophytoplankton groups in 2000 and 2003 (Fig. 7(a)), and also of heterotrophic bacteria in 2003 (Fig. 7(b)).
The PN and YT sections covered broad range of salinity in relation to the distance from the land; the data of these sections were used to examine the relationship between picoplankton and salinity. The heterotrophic bacterial numbers decreased in the water column with higher salinity (Fig. 7(c)), while no relationship could be found between salinity and concentrations of picophytoplankton, with Prochlorococcus disappearing at salinity <29. This revealed that Prochlorococcus, which is commonly more abundant in oligotrophic waters [10,11], could be eliminated in the low-salinity and turbid coastal waters, which is, however, opposite to euryhaline groups such as Synechococcus and picoeukaryotes.
Meanwhile, all picophytoplankton groups showed a negative correlation between abundance and depth (Fig. 7(d)). This may reflect the dependence of those autotrophs upon light. Also, due to the stratification that tends to constrain the biogenic DOM in the surface water , the heterotrophic bacterial number displays a negative correlation with depth as well (Fig. 7(e)).
Nutrient supply (e.g. N and P) is an important factor in maintaining the picoplankton distribution . Outside the Changjiang Estuary, we observed an upwelling of cold and saline waters with high nutrients , similar to the YT section. Other factors such as iron , grazing by ciliates  or other microzooplankton, viral infections and co-sedimentation with organic particles [36,37] may also affect the picoplankton distribution, however, they were not tested in this study.
A positive correlation between heterotrophic bacteria and the autotrophs (i.e. Synechococcus, Prochlorococcus and picoeukaryotes) was found in this study (Fig. 7(f)), suggesting the dependence of bacteria on the substrates produced by the small primary producers.
Different water masses are found in the East China Sea shelf (Fig. 2(b)). Synechococcus appears to be the most abundant population in the areas affected by Taiwan current warm water, Prochlorococcus becomes the dominant picophytoplankton population in the Kuroshio region. The concentrations of picoeukaryotes and heterotrophic bacteria are comparable among different water masses (Fig. 8). The abrupt change in picoplankton cell numbers within a short distance may be typical of the marginal seas , where hydrographic conditions are rather complicated.
Fixation can lead to loss in Prochlorococcus numbers within a month and there is also a notable decrease in abundances of all picoplankton groups after three months, so the samples should be analyzed on board or as soon as possible when back to the laboratory. In 2000 and 2003, the average numbers of Synechococcus, picoeukaryotes and heterotrophic bacteria in the East China Sea were in the range of 105, 104 and 106 cells ml−1, respectively, with a relatively higher abundance in the inner shelf, where nutrient levels are higher due to Changjiang effluent plumes and the coastal upwelling, compared to offshore oligotrophic waters. Prochlorococcus was absent in the near-shore zones partially because of high turbidity and low salinity but became the dominant population among picophytoplankton groups offshore. A positive correlation with water temperature and a negative correlation with depth were found for all four groups, and a negative correlation with salinity for heterotrophic bacteria throughout the PN and YT sections was also seen. Furthermore, picoplankton groups show different distributions in different water masses. The hydrographic conditions to some extent confine the picoplankton biomass.
This study was funded by the Ministry of Science and Technology of PR China (Nos. G1999043705 and 2001CB711004) and by the Shanghai Priority Academic Discipline Project. We thank Dr. D. Vaulot for the help in population discrimination by FCM and Dr. H. Wei for providing the hydrographic data and assistance in field-work. Prof. R. Laanbroek and two anonymous reviewers are acknowledged for their critical comments to original manuscript.