Seasonal and spatial comparisons of phytoplankton growth and mortality rates due to microzooplankton grazing in the northern South China Sea

Seasonal and spatial comparisons of phytoplankton growth and mortality rates due to microzooplankton grazing in the northern South China Sea B. Chen, L. Zheng, B. Huang, S. Song, and H. Liu State Key Laboratory of Marine Environmental Science and Key Laboratory of the Ministry of Education for Coastal and Wetland Ecosystem, Xiamen University, Xiamen, Fujian, China Division of Life Science, Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong, China Key Laboratory of Marine Ecology and Environmental Science, Institute of Oceanology, Chinese Academy of Sciences, Qingdao, Shandong, China


Back Close
Full upon nutrient enrichment (Landry et al., 2000).Nutrients excreted by microzooplankton are particularly important for maintaining low phytoplankton biomass but relatively high growth rates of phytoplankton in some high-nitrate-low-chlorophyll (HNLC) regions (Frost and Franzen, 1992;Landry et al., 1997;Strom et al., 2000).Recently, Landry et al. (2011a) have shown that the rate of phytoplankton biomass grazed by both microzooplankton and mesozooplankton can fully balance phytoplankton growth rate throughout the euphotic zone in the equatorial Pacific.
It is still unclear how the effect of microzooplankton grazing on primary production changes with environmental conditions such as temperature and nutrient supply.The proportion of daily primary production consumed by microzooplankton (m/µ 0 ) is often believed to be greater in oligotrophic waters where phytoplankton with small size are more edible for microzooplankton (Liu et al., 2002a;Strom et al., 2007).It may also increase with increasing temperature because of different temperature coefficients for phytoplankton and microzooplankton growth (Rose and Caron, 2007).
Light also has the potential to decouple m from µ 0 .From the surface of the ocean to the bottom of euphotic zone, light intensity decreases exponentially, which causes a substantial reduction of µ 0 ; while m may not be affected as much as µ 0 (Landry et al., 2011b).As such, microzooplankton should remove a greater proportion of primary production at depth compared with the light-saturated surface waters.
In spite of the above environmental effects, Calbet and Landry (2004) did not find any systematic trends of m/µ 0 along chlorophyll gradients in their analysis on a global dilution dataset.In another analysis using generalized additive models, Chen et al. (2012) found that the combination of temperature and chlorophyll only explains 4 % of the total variation of m/µ 0 .Are there other parameters that were not taken into account in the above analyses but are important in affecting m/µ 0 ?Or are the m/µ 0 ratios intrinsically not able to being predicted by external environmental parameters?With these questions in mind, we conducted a series of microzooplankton grazing experiments in the northern SCS in two cruises, by taking advantage of the highly variable environments in this area.There are relatively few data on microzooplankton Introduction

Conclusions References
Tables Figures

Back Close
Full herbivory in this area compared with primary production data (Liu et al., 2002b;Chen and Chen, 2006).The northern SCS is affected by a number of physical forcings including the continental runoff from the north, China coastal current coming through the Taiwan Strait, and seasonal reversing monsoons.During summer, the southwest monsoon induces clockwise water current circulation in the northern SCS and coastal upwelling over the widened continental shelf (Wong et al., 2007;Gan et al., 2009), while the upper ocean layer forms a large-scale cyclonic circulation under the influence of the northeast monsoon in winter and the nutrient-rich East China Sea coastal water can flow into NSCS through the Taiwan Strait.The most salient seasonal pattern in offshore waters of SCS is the peak of phytoplankton biomass, primary production, and new production during wintertime when mixed layer deepens and nutrients are entrained into the euphotic zone (Liu et al., 2002b;Ning et al., 2004;Chen and Chen, 2006).In summer, the enhanced Pearl River discharge may also induce higher phytoplankton biomass and primary production in the plume area.We test three hypotheses.First, in the oligotrophic basin waters, should microzooplankton grazing remove a greater proportion of primary production than in more eutrophic shelf waters?Second, in the warm summer, should m/µ 0 be greater than in winter?Third, should m/µ 0 be greater at depth than in surface waters?

Material and methods
Dilution experiments (Landry and Hassett, 1982) were conducted at a total of 46 stations during two cruises, one during the summer (18 July to 16 August 2009; 22 stations) and the other in winter (6 January to 30 January 2010; 24 stations) in the northern SCS (Fig. 1).At each station, seawater samples were collected from two depths (1 m and DCM layer) using an acid-washed Niskin bottle attached to a CTD rosette system.
During the winter, a DCM layer did not exist at many stations (Tables S1, S2) and the so-called "DCM layer" was determined as roughly 5 % of surface irradiance.All incubation bottles, tubing and carboys were washed with 10 % HCl and rinsed thoroughly Introduction

Conclusions References
Tables Figures

Back Close
Full with distilled water and ambient seawater before each experiment.Measured amounts of particle-free water, prepared by gravity filtering the seawater through a 0.2 µm filter capsule (Pall Corporation), were first added to 1.2 L polycarbonate bottles, and the bottles were then gently filled with whole seawater to capacity.The filter capsules were soaked in 10 % HCl for more than 2 h before the first use and were washed with diluted acid, distilled water and ambient seawater between each experiment to eliminate possible toxins associated with the capsules (Landry et al., 1995) For Chl a analyses, 300 mL to 1.2 L seawater samples were filtered onto GF/F glassfiber filters under low vacuum.The filters were extracted in 90 % acetone at 4 • C in the dark for 24 h and the Chl a concentrations were measured by the non-acidification method (Welschmeyer, 1994) on a Turner Designs fluorometer (Model No. Trilogy 040).FCM samples were fixed with 0.5 % buffered paraformadehyde and frozen at −80 • C  (Vaulot et al., 1989).Cell abundances of picophytoplankton were enumerated using a Becton-Dickson FACSCalibur cytometer, with different populations distinguished based on side-scattering (SS), orange and red fluorescence (Olson et al., 1993).Yellow-green fluorescent beads (1 µm, Polysciences) were added to the samples as an internal standard.For counting heterotrophic nanoflagellates, the samples were stained with 0.02 % SYBR Green I (Molecular Probes) in the dark under the presence of 30 mmol L potassium citrate at 37 • C for 1 h before analysis (Zubkov et al., 2006).The exact flow rate was calibrated by weighing a tube filled with distilled water before and after running for certain time intervals and the flow rate was estimated as the slope of a linear regression curve between elapsed time and weight differences (Li and Dickie, 2001).
Ciliates and dinoflagellates were preserved by 5 % acidic Lugol's solution at room temperature until analysis.Upon return to the lab, the samples were observed with an inverted microscopy (Leica Dmirb).Cell length and width were sized using the software Simple PCI6.Cellular carbon content of ciliates was calculated from biovolumes using a conversion factor of 0.19 pg C µm −3 (Putt and Stoecker, 1989).Biovolume of dinoflagellates was converted to cell carbon using the equation: pg C cell −1 = 0.76 × volume (µm 3 ) 0.819 , according to Mender-Deuer and Lessard (Menden-Deuer and Lessard, 2000).Only dinoflagellates known to have phagotrophic ability (such as Gyrodinium, Protoperidium) were included in the biomass of microzooplankton.
Assuming an exponential growth model, we calculated the net growth rate (k i ) of phytoplankton in each dilution treatment according to the formula where C i is the Chl a concentration in the i -th treatment bottle at 24 h, D i is the dilution factor (proportion of unfiltered seawater) of the i -th treatment, and C o is the initial Chl a concentration.Estimates of phytoplankton growth rate with nutrient enrichment (µ n ) and mortality rate (m) were derived from Model I linear regressions of net growth rate against dilution factor (Landry and Hassett, 1982).In situ estimates of phytoplankton instantaneous growth rate (µ 0 ) were computed as the sum of m and net growth rate in control bottles without added nutrients.For a few cases of positive slope of the linear regression (negative grazing rates, but not significantly different from zero), we determined m to be zero and µ n to be the average value of the net growth rates of all five dilution treatments with nutrient enrichment (Murrell et al., 2002).
We used FCM-derived estimates of cellular biovolume and fluorescence to correct Chl a estimates of phytoplankton growth rate for pigment photoacclimation.For each experiment, the ratios (R) of cellular red fluorescence to biovolume were calculated for initial and final FCM samples.Corrected phytoplankton growth rates (µ 0 and µ n ) were Introduction

Conclusions References
Tables Figures

Back Close
Full where R i and R f are the initial and final R estimates (Landry et al., 2003;Chen et al., 2009b).m was not affected by changes in cellular pigment contents.
Corresponding seawater temperature, salinity, pressure, nutrient, and Chl a concentrations were also measured.Temperature, salinity, and pressure were determined by Conductivity-Temperature-Depth (CTD) probes.Mixed layer depth (MLD) was defined as the first depth where temperature was 0.2 • C lower than at surface (5 m).Nutrients were measured following standard methods (Parsons et al., 1984).

Results
For identifying spatial patterns, we classify the stations into three groups according to bathymetry: shelf (bottom depth ≤ 100 m), slope (100 m < bottom depth ≤ 2000 m), and basin (bottom depth > 2000 m).Note that although this crude approach neglects the variations of mesoscale features such as river plume and eddies, it provides a straightforward way to show the major cross-shelf gradients.

Temperature, nutrients, and mixed layer depth (MLD)
The background information of physical and chemical parameters is shown in Table 1.Most summer stations were warm, oligotrophic, and stratified, while winter stations were relatively cool, mesotrophic, and well-mixed.Except for a few stations, there were no evident cross-shelf gradients of temperature and nutrient concentration in summer; while in winter, shelf waters were cooler and richer than slope and basin waters.The depth of DCM layer was usually below MLD in summer but shallower than MLD in winter.Introduction

Conclusions References
Tables Figures

Back Close
Full In summer, surface Chl a concentrations (mean = 0.12 and 0.15 mg m −3 , for slope and basin waters, respectively) were significantly lower than those at DCM layers (mean = 0.62 and 0.57 mg m −3 , respectively) in slope and basin waters (paired Wilcoxon-tests, p < 0.05); while this vertical difference was not observed in winter due to more vigorous mixing (median stratification index = 0.041 and 0.012 kg m −4 , in summer and winter, respectively) (Fig. 2).In shelf waters, Chl a concentrations were insignificantly different between surface and DCM waters in both summer and winter (paired Wilcoxon-tests, p < 0.05; Fig. 2a, b).Comparing seasonal differences, in slope and basin waters, surface Chl a concentrations were significantly lower in summer than in winter (mean = 0.55 and 0.61 mg m −3 , for winter slope and basin waters, respectively; Wilcoxon-tests, p < 0.01); while surface Chl a concentrations were insignificantly different between summer in winter in shelf waters (p > 0.05).In both seasons, spatially, there is a decreasing trend of both surface and DCM Chl a concentrations from shelf to deeper stations (Fig. 2).
In spite of the large differences of Chl a concentrations between surface and DCM in summer slope and basin waters, Bz did not differ significantly between the two depths in summer (p > 0.05; Fig. 3).Within each region, B z did not differ significantly between summer and winter, either (p > 0.05).There is a decreasing trend of surface Bz from shelf to basin waters in summer (Wilcoxon-tests, p < 0.01), but not in winter (Fig. 3).
During summer, B z was positively correlated with Chl a in surface waters (Spearman r = 0.46, p < 0.05), but not in DCM waters (p > 0.05).There was no such positive correlation in the winter.The ratio of Bz over Chl a was significantly greater in surface waters than in DCM in the summer (paired Wilcoxon-test, df = 15, p < 0.001), but not in winter.In surface waters, the ratio of B z : Chl a was also significantly higher in the summer than in the winter (Wilcoxon test, p < 0.001), which might be related with both carbon-to-chlorophyll ratios and microzooplankton-to-phytoplankton biomass ratios.Introduction

Conclusions References
Tables Figures

Back Close
Full

Phytoplankton growth and mortality rates due to microzooplankton grazing
The detailed results for each experiment are shown in data appendices Tables S1  and S2.In both seasons, surface µ 0 (mean ± sd: 0.89 ± 0.45 d −1 and 0.61 ± 0.32 d −1 , for summer and winter, respectively) were significantly higher than that in DCM layers (mean ± sd: 0.29 ± 0.34 d −1 and 0.45 ± 0.21 d −1 , for summer and winter, respectively) (paired Wilcoxon tests, p < 0.01) (Fig. 4).Spatially, there were no significant crossshelf trends of µ 0 in summer or winter surface waters.On average, surface µ 0 were significantly higher in summer than in winter in shelf waters (Wilcoxon test, p < 0.05), but were similar in slope and basin waters (Fig. 4).Phytoplankton mortality rates due to microzooplankton grazing (m) averaged 0.49 ± 0.47 d −1 and 0.35 ± 0.21 d −1 (mean ± sd) for summer and winter, respectively, in surface waters and averaged 0.21 ± 0.13 d −1 and 0.34 ± 0.11 d −1 (mean ± sd) for summer and winter, respectively, in DCM waters.m were significantly higher in surface than in DCM layers (Wilcoxon tests, p < 0.05) in summer shelf and slope waters, but not so in basin or during winter (p > 0.05; Fig. 5).There was a decreasing trend of m from shelf to basin waters (p < 0.05) in summer, but not in winter.No differences of surface m could be found between summer and winter; while m at DCM were significantly lower in summer than in winter (Wilcoxon test, p < 0.01).
The high growth rate of phytoplankton in the summer surface waters was consistent with the relative high nutrient limitation index (µ 0 /µ n ) (median = 85.3 % and 94.4 % in Introduction

Conclusions References
Tables Figures

Back Close
Full summer and winter, respectively; Fig. 7).Surface nitrate concentration was not correlated with µ 0 or µ 0 /µ n in surface waters in either season (p > 0.05).
The relatively high µ 0 in the summer surface waters was partially related with high temperature given the positive correlation between temperature and µ 0 in the pooled dataset (Spearman r s = 0.30, p < 0.05).m was positively correlated with temperature during the winter after excluding a lowest value (r s = 0.46, p < 0.05), but were positively correlated with B z (r s = 0.49, p < 0.05) and Chl a (r s = 0.53, p < 0.05) in summer.

Comparisons of rate estimates with previous studies in the northern SCS and in other areas with similar latitude
Before discussing environmental effects on microzooplankton grazing effects on phytoplankton, it is prudent to compare our data with other studies in the same area or similar environments.There are not many studies on microzooplankton grazing in the northern SCS and, if any, the estimates on phytoplankton growth and microzooplankton grazing rates are concentrated in surface waters.One impression arising from browsing the available data is that the rate estimates are quite variable as responding to the complex coastal hydrographic dynamics such as upwelling, typhoons, coastal current and river plume etc.For example, Huang et al. (2011) reported an average phytoplankton growth rate of 1.02 ± 0. Globally, although hundreds of papers have been published estimating microzooplankton grazing rates using the dilution technique (Calbet and Landry, 2004;Chen et al., 2012), there are relatively few studies at similar latitudes (∼ 20 • N) in open ocean waters.The µ 0 and m estimated by Landry et al. (1998) in the Arabian Sea, which is at similar latitudes with ours, are similar with our estimates both in summer (mean growth rate = 0.85 d −1 and mean grazing rate = 0.68 d −1 at surface) and winter (mean growth rate = 0.62 d −1 and mean grazing rate = 0.65 d −1 at surface) (see their Fig. 3).Their rate estimates at low light (5 % surface irradiance) were also similar to ours.Also in the Arabian Sea, the estimates of Edwards et al. (1999) were slightly lower (growth rate ranged from 0.25 d −1 to 1.77 d −1 and grazing rate from 0.15 d −1 to 0.68 d −1 ) but still lied within the normal range.For the Pacific and Atlantic Ocean at similar latitudes, we are not aware of any comprehensive studies on microzooplankton herbivory.It is still difficult to reliably predict m, not even mentioning m/µ 0 , using remotely sensed variables such as temperature and Chl a concentrations.Using a global dataset we compiled previously (Chen et al., 2012), we found that temperature and Chl a concentrations together explained less than 20 % of total variance of m even using the flexible generalized additive modeling (the authors' unpublished data).Predator-prey interactions within the plankton consortium are complex (Peters, 1994;Poulin and Franks, 2010) and it remains to be investigated whether we should strive to develop a better model or whether it is impossible to predict microzooplankton biomass and grazing activity only relying on remotely sensed variables and we should be conservative on the applications of remote sensing on the heterotrophic processes in the ocean (Banse, 2013).

Vertical variation of phytoplankton growth and microzooplankton grazing rate
Comparison of µ 0 between surface and DCM waters confirms our original hypothesis that light extinction greatly diminishes phytoplankton growth rate at DCM layers.Also consistent with the findings by Landry et al. (2011b), we find similar m between surface and DCM layer in winter and therefore microzooplankton grazed a higher proportion of primary production at DCM layer in winter.As the two sampling depths in winter lied within the surface mixed layer at many stations, it is not surprising to find similar microzooplankton community structure and biomass at the two depths in winter (Fig. 3).Although light has been reported to stimulate the grazing activity of some protists (Strom, 2001), this stimulation effect should not be as strong as the light effect on phytoplankton growth rate.
In contrast, the mean m at DCM layer was also lower than at surface in summer shelf and slope waters and m/µ 0 was not different between the two depths.The similar microzooplankton biomass at the two depths suggests that the difference was mainly due to the grazing activity per capita microzooplankton biomass (m/B z ).
The reason for the reduced m/B z at DCM in summer is unclear.As all the experiments were incubated at surface temperatures, it should not be the temperature effect that caused the reduced m/B z at DCM layers.The level of light screening was similarly used in the two cruises so that it is unlikely that the light differences between the two depths caused the lower m/B z at DCM in summer, but not in winter.Our resolution of identifying microzooplankton species is inadequate to address whether microzooplankton community composition differed significantly between surface and DCM in summer.There was no significant difference of the biomass ratio of ciliates and dinoflagellates between the two depths in summer (data not shown).Either, there was no significant difference of average microzooplankton cell size between the two depths in summer.Phytoplankton biomass and average cell size should not cause the reduced m/B z at DCM in summer because phytoplankton biomass and average cell size was similar Introduction

Conclusions References
Tables Figures

Back Close
Full between summer and winter at DCM.This problem should be further tested in future studies.It also calls for the attention against extrapolating surface rate estimates to deep depths.

Spatial variations of phytoplankton growth and microzooplankton grazing rates
Contrary to our original hypothesis, we observed a decreasing trend of m/µ 0 from the eutrophic shelf waters to the oligotrophic basin waters in summer surface waters, which is partially related with increasing microzooplankton biomass with increasing Chl a concentration.This suggests that microzooplankton biomass instead of phytoplankton size structure is the principle factor determining m.Sherr and Sherr (2007) pointed out that heterotrophic dinoflagellates, which have a variety of feeding mechanisms (Jeong, 1999) and can feed on prey equal or larger than their own size (Hansen et al., 1994), tend to dominate in high Chl waters.
One difference of our study to the high latitude study by Liu et al. (2002a) is that we did not observe a positive correlation between Chl a concentration and phytoplankton growth rate (i.e., µ 0 was not lower in oligotrophic waters), which may cause the decreasing trend of m/µ 0 from shelf to basin waters.Similar to observations in subtropical and equatorial Pacific (Laws et al., 1987;Landry et al., 2011b), we find high phytoplankton growth rates (> 0.5 d −1 ) in the basin surface waters of the SCS especially in the oligotrophic summer, which are probably sustained by grazer nutrient excretion and nitrogen fixation.SCS is well known for the occurrences of internal waves and typhoons (Chen et al., 2009c), which can disturb the stratified water column and periodically inject the nutrients into the euphotic zone from below.The effect of nutrient enrichment on µ 0 is small, suggesting that phytoplankton were not experiencing severe nutrient limitation at this time.Marra and Barber (2005) and Behrenfeld (2010) suggested that the key factor regulating the variations of phytoplankton biomass in the Arabian Sea and the North Atlantic is likely the changing grazing effect induced by the mixing process, instead of 16017 Introduction

Conclusions References
Tables Figures

Back Close
Full bottom-up factors such as nutrients or light.When vertical mixing occurs induced either by upwelling or winter surface cooling, the particle-poor subsurface waters dilute the surface water within the euphotic zone, acting as a natural "dilution" experiment.The grazer biomass and grazing impact on phytoplankton decreases and, as a consequence, net growth rate of phytoplankton becomes positive and phytoplankton biomass accumulates.Whether phytoplankton growth rate (µ 0 ) increases or not after the mixing event is not a key issue here.Our data partially support this hypothesis.The similar µ 0 between summer and winter in basin waters suggests that the elevated nutrient levels induced by winter mixing do not substantially increase phytoplankton growth rate.
It should be the relaxed grazing pressure induced by natural mixing event that leads to net positive growth of phytoplankton during the progression from summer to winter.However, at present, we do not have sufficient time-series data of phytoplankton biomass, µ 0 and m to fully validate this hypothesis.

Temperature effects on microzooplankton grazing effect
While it is well founded in theory that the growth rate of phytoplankton should increase more slowly with temperature than microzooplankton growth and grazing rates (Lopez-Urrutia et al., 2006;Rose and Caron, 2007;Lopez-Urrutia, 2008), it is difficult to disentangle the individual effect of temperature in the overall grazing impact.For example, comparing the surface basin waters between summer and winter, m/µ 0 was lower in the warmer summer than the colder winter (Fig. 7) due to the negative correlation between temperature and grazer biomass.
The temperature effect is supposed to be more pronounced in eutrophic shelf waters because seasonal variations of temperature are greater than in basin waters and also because the negative correlation between temperature and grazer biomass is weaker.When we calculate the temperature coefficient of m only for shelf surface waters by linearly regressing ln m against 1/kT, the mean activation energy is 0.70 eV, close to the global average 0.65 eV (Chen et al., 2012).In comparison, the mean activation

BGD Introduction
Full energy of m for all the data is only 0.28 eV.These calculations suggest that temperature effects on microzooplankton grazing are more salient in shelf waters.
In summary, we have conducted a comprehensive study on microzooplankton herbivory in the northern SCS.Although microzooplankton herbivory is an important loss pathway of primary production, we still do not have sufficient measurements in the ocean particularly in the lower part of the euphotic zone (Landry et al., 2011b).As a consequence, there is still no widely accepted theory on microzooplankton grazing that can easily fit to field data.Although the global average proportion of primary production grazed by microzooplankton is estimated as from 60 % to 80 %, the real ratio of m/µ 0 can range from 0 to 100 % with little predictability (Calbet and Landry, 2004;Chen et al., 2012).While primary production has been mapped at global scales using remote sensing techniques, the estimates of microzooplankton grazing rate are largely scattered.But these estimates are essential for understanding the dynamics of phytoplankton biomass in the ocean (Banse, 2013).Clearly, plankton ecologists need more accurate measurements in the ocean and also need to develop sophisticated theories that can capture the essence of microzooplankton grazing and can do relatively well in prediction.

Supplementary material related to this article is available
Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | calculated as µ Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | 3.2 Chl a and microzooplankton biomass (B z ) Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | 27 d −1 and an average microzooplankton grazing rate of 0.85 ± 0.37 d −1 in upwelling regions of northeastern SCS during summertime.While in non-upwelling regions, the rate estimates lowered to 0.51 ± 0.05 d −1 and 0.50 ± 0.17 d −1 for phytoplankton growth and microzooplankton grazing, respectively.Zhou et al. (2011) also estimated phytoplankton growth and microzooplankton grazing rates in the northeastern SCS after passage of a typhoon.They seemed had sampled a post-bloom phase as many of their experiments demonstrated negative phytoplankton growth rates and the microzooplankton grazing rates were highly variable.Chen et al. (2009a) and Lie Discussion Paper | Discussion Paper | Discussion Paper | and Wong (2010) have reported high phytoplankton growth (> 1.5 d −1 ) and microzooplankton grazing rates (> 1 d −1 ) in Hong Kong nearshore waters during summertime, which are more eutrophic than most of our sampling stations.Su et al. (2007) also reported high phytoplankton growth and microzooplankton grazing rates at a coastal station near Hong Kong.Their estimates (∼ 0.1 d −1 ) at 75 m of other 4 basin stations are similar to our estimates at DCM layers of basin waters in summer.
Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Fig.1

Fig. 2 .Fig. 4 .
Fig. 2. Boxplots of Chl a (mg m −3 ) in surface and DCM waters.The line through the middle of the box shows the median.The outer edges of the box correspond to the 25th and 75th percentiles, and the "whiskers" to the 10th and 90th percentiles.The dots represent extreme values.