Physiological responses of coastal and oceanic diatoms to diurnal fluctuations in seawater carbonate chemistry under two CO 2 concentrations

Diel and seasonal fluctuations in seawater carbonate chemistry are common in coastal waters, while in the open-ocean carbonate chemistry is much less variable. In both of these environments, ongoing ocean acidification is being superimposed on the natural dynamics of the carbonate buffer system to influence the physiology of phytoplankton. Here, we show that a coastal Thalassiosira weissflogii isolate and an oceanic diatom, Thalassiosira oceanica, respond differentially to diurnal fluctuating carbonate chemistry in current and ocean acidification (OA) scenarios. A fluctuating carbonate chemistry regime showed positive or negligible effects on physiological performance of the coastal species. In contrast, the oceanic species was significantly negatively affected. The fluctuating regime reduced photosynthetic oxygen evolution rates and enhanced dark respiration rates of T. oceanica under ambient CO2 concentration, while in the OA scenario the fluctuating regime depressed its growth rate, chlorophyll a content, and elemental production rates. These contrasting physiological performances of coastal and oceanic diatoms indicate that they differ in the ability to cope with dynamic pCO2. We propose that, in addition to the ability to cope with light, nutrient, and predation pressure, the ability to acclimate to dynamic carbonate chemistry may act as one determinant of the spatial distribution of diatom species. Habitat-relevant diurnal changes in seawater carbonate chemistry can interact with OA to differentially affect diatoms in coastal and pelagic waters.


Introduction
Anthropogenic emissions of carbon dioxide (CO 2 ) since the Industrial Revolution have increased atmospheric pCO 2 levels by 40 % (Howes et al., 2015), mainly due to burning of fossil fuels and land use changes (Ciais et al., 2014).The oceans absorb about 30 % of the CO 2 emitted by human activities (Sabine et al., 2004), leading to decreases in pH, concentration of carbonate ions, and saturation state of calcium carbonate, along with increases in the concentrations of aqueous CO 2 and bicarbonate (i.e., ocean acidification).The global surface ocean mean pH has already decreased by about 0.1 units since the Industrial Revolution (Orr et al., 2005;Doney, 2010), and a further decrease of 0.3-0.4 units is expected to happen by 2100 under the "business-as-usual" scenario (Orr et al., 2005;Gattuso et al., 2015).
For marine organisms, the reduced seawater mean pH caused by ocean acidification (OA) could be detectable on a timescale of years to decades, while striking fluctuations in coastal seawater carbonate chemistry may occur over much shorter timescales in current and OA scenarios.The coastal zone plays a critical role in biogeochemical cycles and experiences great variability of physical and chemical factors (Drupp et al., 2011).In addition, it is the area most impacted by anthropogenic pressures (Gattuso et al., 1998).Carbonate chemistry in coastal seawater is affected by multiple drivers in addition to atmospheric CO 2 dissolution, such as tidal cycles (Dai et al., 2009;Jiang et al., 2011;Wang et al., 2014), upwelling (Feely et al., 2008;Capone and Hutchins 2013), watershed processes, wind forcing (Drupp et al., 2011), anthropogenic nutrient inputs, aquaculture ac-F.Li et al.: Ocean acidification influences diatoms under fluctuating pH tivities, and changes in ecosystem structure and metabolism (Duarte et al., 2013;Waldbusser and Salisbury 2014).Due to high biomass and sufficient or excess nutrients in coastal waters, biological activities alter pCO 2 , resulting in a diel cycle of pH.The diel range of pH variation in some coastal ecosystems can be greater than 1 pH unit (Duarte et al., 2013), which corresponds to a 900 % change in H + concentration.
During a diurnal cycle, organisms in coastal areas could experience pH values that may be lower than the projected value for the surface ocean in the year 2100 (Hofmann et al., 2011;Hurd et al., 2011;Waldbusser and Salisbury, 2014).In contrast, pH in the open ocean is relatively stable, with a variation range of only ∼ 0.024 over a month (Hofmann et al., 2011).Considering the lower buffering capacity in the OA scenario, pH variability would increase in both coastal and oceanic waters (Egleston et al., 2010;Cai et al., 2011;Denman et al., 2011;Wang et al., 2013).The amplitude of pH variation in coastal water will be larger than in oceanic water due to the presence of multiple drivers (Waldbusser and Salisbury, 2014).For instance, biological activities could increase variation in pH up to 40 % compared to the present extent of variation under elevated pCO 2 conditions in coastal waters (Egleston et al., 2010).
Responses of fish (Dixon, 2014), gastropods (Onitsuka et al., 2014), oysters (Keppel, 2015), mussels (Frieder et al., 2014), coral (Dufault et al., 2012;Comeau et al., 2014), canopy-forming kelp (Britton et al., 2016), and coralline algae (Gao et al., 1993;Cornwall et al., 2013;Noisette et al., 2013;Johnson et al., 2014) to diurnally fluctuating pCO 2 /pH have been studied recently.Dufault et al. (2012) hypothesized that storage of dissolved inorganic carbon during the nighttime high-pCO 2 period fueled daytime calcification (and perhaps photosynthesis), resulting in higher calcification and survival rate of coral recruits.Thus, it appears that some marine organisms may benefit from pCO 2 fluctuations.In spite of this body of literature, the responses of marine phytoplankton to fluctuating pH/pCO 2 are still unclear.To our knowledge, only one study has addressed the responses of the marine green alga Ostreococcus to fluctuating pCO 2 (Schaum et al., 2016).However, how CO 2 variability affects other major marine phytoplankton groups over either the short or long term remains unknown.
Coastal and open-ocean species are distinguished by habitat-related differences in cell size, nutrient utilization (Glibert andRay, 1990), photosynthetic architecture (Strzepek andHarrison, 2004), and photosynthetic performance (Lavaud et al., 2007;Li et al., 2011;Liu and Qiu, 2012).Our study was intended to understand whether coastal and oceanic species also differ in their capacity to respond to fluctuating carbonate chemistry.A coastal Thalassiosira weissflogii isolate and an oceanic diatom, Thalassiosira oceanica, were used in the present study.We manipulated pCO 2 to mimic diurnally fluctuating carbonate chemistry and hypothesized that coastal diatoms would show better physiological performance under fluctuating carbonate chemistry than oceanic ones, a difference that could potentially be a key factor influencing the geographical distribution of diatoms.

Cultures and experimental setup
Thalassiosira weissflogii (CCMP 1336, isolated from coastal Long Island, New York, USA, in 1956) and Thalassiosira oceanica (CCMP 1005, isolated from the Sargasso Sea in 1958) were incubated in Aquil medium (Sunda et al., 2005).Triplicate cultures (incubated in 1 L autoclaved Erlenmeyer flasks) were used for each treatment, illuminated by cool white fluorescent light at an intensity of 115 µmol photons m −2 s −1 .Cultures were maintained at 20 • C with a 12 h : 12 h light and dark cycle.Cells were maintained at exponential growth phase with maximal concentration < 1.1 × 10 4 mL −1 (T.weissflogii) or 3.5 × 10 4 mL −1 (T.oceanica) in semi-continuous cultures (cultures were diluted every 24 h at 6 h after the onset of light).
T. weissflogii and T. oceanica were acclimated to four treatments: (1) steady carbonate chemistry at ambient pCO 2 level (LCs); (2) diurnally carbonate chemistry fluctuating around ambient pCO 2 level (LCf); (3) steady carbonate chemistry at elevated pCO 2 level (HCs); and (4) diurnally carbonate chemistry fluctuating around elevated pCO 2 level (HCf) for 15 generations before sampling.Steady regimes were bubbled with ambient air (400 ± 15µatm, LCs) or elevated (1005 ± 40 µatm, HCs) pCO 2 , which was automatically achieved by mixing air/CO 2 with a CO 2 enricher (CE100B, RuiHua).The fluctuating regimes were obtained by changing the CO 2 partial pressure every 12 h.Cells were aerated with air of low pCO 2 (i.e., 0 or 557 ± 15 µatm for LCf and HCf, respectively) during the photoperiod; the aeration was changed to high pCO 2 (i.e., 870 ± 19 or 1949 ± 35 µatm for LCf and HCf, respectively) at the beginning of the dark period.Measurements showed that pH gradually increased and decreased, similar to a natural diurnal cycle (see Results).Since pH increased quickly in the first few hours of the photoperiod, the aeration rates were adjusted to make sure the fluctuating regimes reached similar pH values with corresponding steady regimes in the middle of photoperiod and reached target values at the end of photoperiod.The steady regimes were aerated with stable pCO 2 air at the same flow rate as the fluctuating regimes.The pH was measured every 1.5 h by a pH meter (Orion 2 STAR, Thermo Scientific) calibrated with standard National Bureau of Standards (NBS) buffers.Samples for total alkalinity (TA) measurement were poisoned with a saturated solution of mercuric chloride after filtration.TA was determined by Gran acidimetric titration with a TA analyzer (AS-ALK1+, Apollo SciTech).Certified reference materials obtained from A. G. Dickson at the Scripps Institution of Oceanography were used to assure the accuracy of the TA measurement.TA and pH were applied to CO2SYS software to calculate other carbonate chemistry parameters (Table 1).Subsamples for measurement of physiological parameters were always taken in the middle of the photoperiod (6 h after the onset of light), unless otherwise noted.

Growth rate and chlorophyll a content
Cell concentration and mean cell size were measured by a Coulter particle count and size analyzer (Z2, Beckman Coulter).Specific growth rate was calculated according to the equation µ = (lnN 1 − lnN 0 ) / (t 1 − t 0 ), in which N 1 and N 0 represent cell concentrations at t 1 and t 0 .For the chlorophyll a content determination, samples were filtered onto GF/F filters (25 mm, Whatman) and extracted overnight at 4 • C in absolute methanol before centrifugation.The supernatants were analyzed by a UV-VIS spectrophotometer (DU800, Beckman Coulter) and the chlorophyll a content was calculated according to the equation of Ritchie (2006).

Elemental composition and production rate
Samples for measuring particulate organic carbon (POC) and nitrogen (PON) were filtered onto pre-combusted (450 • C for 6 h) GF/F filters (25 mm, Whatman).Filters were treated using HCl fumes to remove any inorganic carbon and dried before analysis on a CHNS/O analyzer (2400 Series II, PerkinElmer).Polycarbonate filters (1.2 µm pore size) were used to determine biogenic silica (BSi) by the spectrophotometric method of Brzezinski and Nelson (1995).Production rates of POC, PON, and BSi were calculated by multiplying cellular content by specific growth rate.

Chlorophyll a fluorescence
The photochemical parameters were determined using a xenon-pulse amplitude modulated fluorometer (Xe-PAM, Walz).Effective photochemical quantum yields were determined according to the equation of Genty et al. (1989) for light-adapted samples, where F m indicates maximum chlorophyll fluorescence of light-adapted samples and F t indicates steady chlorophyll fluorescence of light-adapted samples.Non-photochemical quenching (NPQ) was calculated as NPQ = (F m − F m )/F m , where F m indicates maximum chlorophyll fluorescence of dark-adapted samples.PSII and NPQ were measured under actinic light intensity (∼ 156 µmol photons m −2 s −1 ) similar to culture light level after 10 min of dark adaptation.Given the changing carbonate chemistry over a diurnal cycle, PSII and NPQ were determined at three time points: 0.5, 6, and 11.5 h after the onset of light.

Photosynthetic oxygen evolution and dark respiration rates
Net photosynthetic oxygen evolution and dark respiration rates were determined using a Clark-type oxygen electrode (Oxygraph, Hansatech) at the experimental temperature.Oxygen evolution rates were measured under 115 µmol photons m −2 s −1 at the same three time points as mentioned above.Oxygen consumption rates were measured in the middle of photoperiod (6 h after the onset of light), when the steady and fluctuating regimes reached similar pH values.Samples were gently filtered (< 0.02 MPa) onto 47 mm cellulose acetate membranes, and then re-suspended into 20 mmol L −1 Tris-buffered medium.The re-suspended cells were injected into an oxygen electrode chamber equipped with a magnetic stirrer.Rates of oxygen evolution and consumption were derived from the linear portion of the slope of the oxygen record (∼ 10 min per sample).The pH values of Tris-buffered media were pre-adjusted to their corresponding culture media values.That is, pH values of Tris-buffered media of the three time points in the LCf treatment were 7.84, 8.14, and 8.35, and those in the HCf treatment were 7.54, 7.80, and 8.06.Values in the LCs and HCs treatments were set to 8.14 or 7.80 for all three time points, respectively.

Statistical analyses
Data were analyzed by a two-way analysis of variance (ANOVA) with pCO 2 level and pCO 2 variability classed as factors in the model, each with two levels (400 ± 15µatm and 1005 ± 40 µatm, two levels of pCO 2 ; steady and fluctuating regimes for two levels of pCO 2 variability, respectively).The interaction of the two factors was also included in the model.All data used for ANOVA analysis were tested for normality (Shapiro-Wilk test) and homogeneity of variances (Levene test).When p values were under 0.05, the post hoc Duncan test was used to determine the differences between treatments.All data are reported as the mean value of triplicate cultures ± standard deviation (SD).

Variation of pH in experimental regimes
The variation ranges of pH in the LCf and HCf treatments were 0.52 ± 0.03 and 0.53 ± 0.03, respectively.At the beginning of the photoperiod, pH of the LCf regime was 7.84 ± 0.02 (Fig. 1), and then it increased to 8.15 ± 0.03 at 6 h after the onset of light, similar to the value of the LCs regime (8.13 ± 0.02).The pH value of the LCf regime reached 8.35 ± 0.02 at 12 h after the onset of light, and then decreased to 7.84 ± 0.02.For the HCf regime, pH ranged from 7.54 ± 0.01 to 8.06 ± 0.02 and reached 7.82 ± 0.01 at

Specific growth rate and mean cell size
Growth rates of T. weissflogii were not influenced by diurnally fluctuating carbonate chemistry in either the current or the OA scenario (Fig. 2a).No effects of OA on growth rate of T. weissflogii were detected.There were no differences in growth rates between the steady and fluctuating regimes for T. oceanica under the ambient pCO 2 condition (Fig. 2b).However, the fluctuating regime reduced its growth rate by 9 % under the elevated pCO 2 condition.OA influenced the growth rate of T. oceanica, with rates of HCs cells being 16 % lower than those of LCs cells.A significant interaction between pCO 2 level and pCO 2 variability on growth rate of T. oceanica was found.
Mean cell sizes were not affected by the fluctuating treatment under either ambient or elevated pCO 2 conditions in T. weissflogii (Table 2).T. oceanica cells showed minor but significant changes in cell size in the fluctuating treatments.Cells in the LCf treatment cells were 1.2 % larger than LCs cells, while HCf cells were 1.4 % smaller than cells in the corresponding steady treatments, resulting in a significant interaction between pCO 2 level and pCO 2 variability.

Chlorophyll a content and elemental composition
Chlorophyll a contents of T. weissflogii in the four treatments were not significantly different.For T. oceanica, the fluctuating regime did not influence chlorophyll a content under ambient CO 2 level.However, in the HCf treatment chlorophyll a content decreased by 24 % compared to the steady regime (Table 2).A significant interaction between pCO 2 level and pCO 2 variability on chlorophyll a content of T. oceanica was found.
POC and PON quotas of both species were elevated in the OA scenario in both the steady and fluctuating regimes, relative to present-day pCO 2 levels (Table 2).However, no effects of the fluctuating regime on cellular POC and PON contents were detected in either species compared to the steady treatments.The only exception was that POC increased by 9 % in the LCf treatment relative to the LCs treatment for T. weissflogii.Generally, elevated pCO 2 and the fluctuating regime showed no effects on BSi quota of either species, besides a slight decrease in the HCf treatment relative to that of the HCs treatment for T. weissflogii.
The fluctuating regime increased the POC production rate of T. weissflogii at both ambient and elevated pCO 2 levels, but had no effects on other elemental production rates of this species.By contrast, the fluctuating regime decreased all of the elemental production rates in the OA scenario for T. oceanica (Fig. 3).Significant interactions between pCO 2  0.16 ± 0.01 a 0.14 ± 0.01 b 0.12 ± 0.01 c 0.10 ± 0.01 c 0.24 ± 0.02 A 0.20 ± 0.02 B 0.14 ± 0.01 C 0.13 ± 0.01 C R : P (fmol cell −1 h −1 : 0.08 ± 0.01 a 0.08 ± 0.02 a 0.13 ± 0.03 b 0.14 ± 0.03 b 0.27 ± 0.07 A 0.55 ± 0.12 B 0.29 ± 0.06 A 0.39 ± 0.04 A fmol cell −1 h −1 ) level and pCO 2 variability on elemental production rates of T. oceanica were found.Slight but significant decreases in the Si : C ratio in the fluctuating regime compared to the steady regime were found at ambient pCO 2 for both species (Table 2).The Si : C ratio of T. weissflogii was lower in the OA scenario, and C : N ratios of T. weissflogii were not significant different among the four treatments.For T. oceanica, cells showed lower C : N and Si : C ratios at elevated pCO 2 relative to cells grown at ambient pCO 2 .

Chlorophyll a fluorescence
The effective photochemical quantum yields of both species varied little at different time points, ranging from 0.54 ± 0.03 to 0.61 ± 0.03 among treatments (Table 3).Fluctuating regimes scarcely influenced PSII of either species.The only exception was that PSII of HCf decreased by 8 % relative to that of the HCs for T. weissflogii at 0.5 h after the onset of light.Elevated pCO 2 decreased PSII by 3 and 2 % for T. weissflogii at 6 and 11.5 h after the onset of light, respectively.T. oceanica cells under elevated pCO 2 showed 5 and 7 % lower PSII compared to those under ambient pCO 2 at 6 and 11.5 h after the onset of light, respectively.NPQ under culture light intensity ranged from 0.06 ± 0.01 to 0.23 ± 0.05 at different time points.No detectable effects of the fluctuating regime on NPQ of either species were found, with the exceptions of HCf cells of T. weissflogii at 0.5 h after the onwww.biogeosciences.net/13/6247/2016/Biogeosciences, 13, 6247-6259, 2016

Photosynthetic oxygen evolution and dark respiration rates
Chlorophyll-normalized net oxygen evolution rates of these two species ranged from 0.39 ± 0.07 to 0.55 ± 0.07 µmol O 2 µg Chl a −1 h −1 at 6 h after the onset of light.Neither elevated pCO 2 nor the fluctuating regime showed detectable effects on oxygen evolution rates per chlorophyll of T. weissflogii (Fig. 4a), while T. oceanica cells under the LCf treatment showed a 29 % lower chlorophyll-normalized net oxygen evolution rate relative to the LCs cells (Fig. 4b).A significant interaction between pCO 2 level and pCO 2 variability on chlorophyll-normalized net oxygen evolution rate of T. oceanica was found.Both species, regardless of treatment, showed a similar diurnal rhythm of photosynthetic oxygen evolution: oxygen evolution rates reached the highest values at 6 h after the onset of light (Fig. 4c, d).For T. weissflogii, effects of fluctuating pCO 2 on net oxygen evolution per cell were only observed at 6 h after the onset of light for cells at the ambient pCO 2 level, with 7 % lower rates in the LCf treatment than in the LCs (Fig. 4c).These effects were more obvious for T. oceanica cells.T. oceanica cells in the steady regime under elevated pCO 2 evolved oxygen at 65 % higher cell-specific rates than those in the fluctuating regime at 0.5 h after the onset of light (Fig. 4d).At 11.5 h after the onset of light, LCs cells of T. oceanica showed 41 % higher net oxygen evolution rates per cell than LCf cells.There were no differences in photosynthetic oxygen evolution rates between HCs and HCf cells 11.5 h after the onset of light.
Elevated pCO 2 increased dark respiration of T. weissflogii by 57 % compared to that at ambient pCO 2 level, while the fluctuating regime had no detectable effect (Fig. 5a).In contrast, dark respiration rates of T. oceanica were stimulated by 60 % for cells under the fluctuating regime compared to steady one at ambient pCO 2 level (Fig. 5b), while no effects of the fluctuating regime at elevated pCO 2 were observed.Dark respiration rates of T. oceanica were similar in www.biogeosciences.net/13/6247/2016/Biogeosciences, 13, 6247-6259, 2016 the steady regimes of ambient and elevated pCO 2 levels.The respiration to net photosynthesis (R : P) ratios for T. weissflogii under elevated pCO 2 was higher than at ambient pCO 2 by 73 %, while no effects of the fluctuating regime were detected.R : P ratios for T. oceanica cells were higher by 104 % in the fluctuating regime than for cells in the corresponding steady regime at ambient pCO 2 level (Table 2).

Discussion
Both species were influenced by elevated pCO 2 in several ways, while they responded differentially to the fluctuating regime.In general, for the coastal diatom T. weissflogii, the fluctuating pCO 2 regime showed either positive (POC cellular quota and production rate) or no obvious effects on its physiological performance.In contrast, the oceanic diatom T. oceanica was significantly negatively affected by the diurnal variation of carbonate chemistry.The fluctuating regime reduced photosynthetic oxygen evolution rates and enhanced dark respiration rates under ambient pCO 2 concentration, while in the OA scenario, the fluctuating regime depressed its growth rate, chlorophyll a content, and elemental production rates (which were caused by decreased growth rates).
OA depressed the growth of T. oceanica, consistent with results of a previous study (King et al., 2015), which showed a similar decrease (19 %) to the present study (16 %).No detectable effects of OA on growth of T. weissflogii were found, as reported by previous studies (Burkhardt et al., 1999;Shi et al., 2009;Reinfelder, 2012;King et al., 2015;Passow and Laws, 2015;Taucher et al., 2015).However, the growth responses of diatoms have also been shown to be affected by interactions between OA and other abiotic factors.For in-stance, the energy saved from active inorganic carbon acquisition mechanisms due to increased availability of CO 2 under OA conditions enhanced the growth of diatoms when daytime mean light level was lower than 22-36 % of sea surface solar light intensity.However, growth under OA condition decreased when light exceeded 25-42 % of incident irradiance (Gao et al., 2012).OA reduced the growth rate of T. weissflogii under light and temperature stress, but no effects of OA were detected in the absence of temperature stress (Passow and Laws, 2015).Consequently, it appears that effects of OA on phytoplankton species could be regionspecific, depending on the local interactions with other abiotic factors.
The silicified cell walls of diatoms act as mechanical protection to resist grazers (Hamm et al., 2003) and also have potential roles in photoprotection (Raven and Waite, 2004), as well as promotion of catalysis by extracellular carbonic anhydrase (Milligan and Morel 2002).The Si : C ratio of both species decreased under the elevated pCO 2 condition, in accordance with results of Tatters et al. (2012) and Mejia et al. (2013).This decreased ratio indicates that the tested species may fix more carbon per silicon assimilated in the OA scenario than under the ambient pCO 2 condition, thus having implications for changes in local and global carbon and silicon budgets.More carbon may be fixed per diatom cell without changes in silicon quota in the OA scenario, and thus the tested species may contribute more to primary production in the ecosystem, especially in Si-limited waters, in the future oceans.However, diatom silicification is under a complex set of controls.For instance, limitation by other nutrients, such as iron (Hutchins and Bruland, 1998) and nitrogen (Flynn and Martin-Jézéquel, 2000), may act to increase Si quotas and Si : C ratio.
Bicarbonate utilization has been suggested to be a general characteristic of marine diatoms, through direct transport or conversion by extracellular carbonic anhydrase (eCA), while the fraction of direct bicarbonate transport and eCA expression varies among species (Martin and Tortell, 2008).Pathways that can utilize HCO − 3 and provide CO 2 for RuBisCo through C 4 (Reinfelder et al., 2000) or C 3 -C 4 intermediate photosynthesis (Roberts et al., 2007) have been suggested for T. weissflogii.This species takes up both CO 2 and HCO − 3 at a similar rate and has the ability to adjust uptake rates to cope with a wide range of inorganic carbon supplies (Burkhardt et al., 2001).Moreover, T. weissflogii has a markedly higher fraction of direct bicarbonate transport and apparent eCA activity than T. oceanica (Martin and Tortell, 2008), which may facilitate their inorganic carbon transport and uptake.In this study, T. oceanica showed significantly lower oxygen evolution rates in the LCf treatment than in the LCs treatment at 11.5 h after the onset of light, when the highest pH and lowest CO 2 was reached.In contrast, no effects of the fluctuating regime on oxygen evolution rates of T. weissflogii were found at this time point.Thus, T. weissflogii cells were more tolerant of the high pH and low CO 2 period under fluctuating carbonate chemistry than T. oceanica.
Under the fluctuating regime, T. oceanica showed higher respiration rates in the current scenario than under the corresponding steady regime.As with the successful compromise between iron requirements and capacity to acclimate to dynamic light regimes in T. oceanica cells (Strzepek and Harrison 2004), this oceanic diatom may also have evolved to acclimate to fluctuating carbonate chemistry in a different way compared with the coastal diatom.The higher respiration rate under the fluctuating regime in the current scenario may imply that this species needs more energy for maintaining its intracellular acid-base balance under dynamic extracellular pH conditions, as dark respiration provides energy for growth and metabolic processes (Raven and Beardall, 2005).Diatoms were shown to exhibit circadian variations in photosynthesis and respiration (Weger et al., 1989;Chen and Gao, 2004).Thus, dark respiration might show different pattern at other time points, as photosynthetic oxygen evolution did.T. oceanica cells showed significantly higher R : P ratios than T. weissflogii, especially in the fluctuating regime at ambient pCO 2 , and the ratios were within previously reported ranges in diatoms (Geider and Osborne, 1989).The higher R : P ratio indicated greater proportions of photosynthetic fixed carbon and associated energy were used for growth, biosynthesis, and maintaining intracellular homeostasis in the oceanic species.Moreover, the fluctuating regime reduced the production rate of organic matter by T. oceanica at elevated pCO 2 .Depressed biomass build-up has also been found under dynamic light regimes (Wagner et al., 2006;Shatwell et al., 2012;Hoppe et al., 2015).Together with our results, this may imply that organisms that are sensitive to fluctuating abiotic factors maintain intracellular homeostasis under dynamic environments of light or pCO 2 at the expense of reduced biomass production.
Either positive (POC production rate) or no obvious effects of the fluctuating regime on biomass production were found in the coastal species T. weissflogii.Coastal calcifying organisms have shown the ability to achieve homeostasis within critical tissues to facilitate calcification under dynamic pH / pCO 2 condition, and this was suggested to be associated with diurnal and seasonal pH fluctuations in coastal waters (Hendriks et al., 2015).Thus, some organisms could take advantage of the fluctuating carbonate system regime to mitigate the negative effects of ocean acidification on physiological performance.For instance, growth and calcification of corals can benefit from oscillatory pCO 2 (Dufault et al., 2012;Comeau et al., 2014).Organisms like T. weissflogii whose physiological performance were enhanced or unaltered under dynamic carbonate chemistry conditions thus could be at a distinct advantage in competing with species that showed negative responses to this condition (such as T. oceanica in the present study).The differential responses of the tested two species to the fluctuating carbonate chemistry may be partially attributed to the differences in cell size.The differences in carbonate chemistry and pH between the bulk medium and the exterior surface of marine organisms increase as cell size increases (Flynn et al., 2012).Thus, the larger species, T. weissflogii, theoretically possesses higher adaptability to cope with the varied carbonate chemistry and pH, as it is frequently encountered in the natural coastal waters and their exterior surfaces.Schaum et al. (2016) found that short-term plastic responses to high pCO 2 disappeared in a green microalgae after extended experimental evolution at high pCO 2 , particularly in fluctuating pCO 2 treatments.Whether a similar phenomenon may be operative in other algal groups, such as diatoms following exposures to high, fluctuating pCO 2 that are longer than those we employed, is currently unknown.However, it is notable that growth rates and competitive abilities of diatoms of a natural community showed little change following 1 year of conditioning at two pCO 2 levels and three temperatures, relative to the results of a short-term experiment conducted on the original collected community (Tatters et al., 2013).Regardless of the responses of cell physiology to different timescales of changes in pCO 2 concentrations, it is a significant observation that the fluctuating regime reduced the production rate of organic matter in T. oceanica at elevated pCO 2 (which were caused by decreased growth rates).
The effect of the fluctuating regime on T. oceanica was different in the current and OA scenarios.Under elevated rather than current pCO 2 conditions, fluctuating carbonate chemistry decreased pigment content and the production rate of organic matter.Although elevated CO 2 mitigated the negative effects of the fluctuating regime on photosynthetic oxygen evolution rates of T. oceanica cells under ambient pCO 2 condition, the effect of the fluctuating regime under elevated F. Li et al.: Ocean acidification influences diatoms under fluctuating pH pCO 2 tended to be negative, resulting in a decreased growth rate compared to the steady regime.The diurnal pH variation range (∼ 0.5 units) used in the present study is realistic for coastal ecosystems, like upwelling regions (Hofmann et al., 2011), kelp forests (Cornwall et al., 2013), coastal coral reefs (Wang et al., 2014), and tide pools (Morris and Taylor, 1983).In contrast, pH in the open ocean is relatively stable, with a variation range of only ∼ 0.024 over a month (Hofmann et al., 2011).The same amplitude of pH variation was set in the current and elevated pCO 2 scenarios in the present study.Buffering capacity will decrease as the dissolved inorganic carbon in both coastal and oceanic seawater increases under projected elevated pCO 2 conditions (Egleston et al., 2010;Cai et al., 2011;Denman et al., 2011;Wang et al., 2013).For instance, the Revelle factor increased from 10.6 ± 0.2 to 15.0 ± 0.2 (a higher Revelle factor indicates a lower buffer capacity) when pCO 2 increased from the ambient to the elevated level in the present study.The increase amplitude of pH variation in coastal water will be more apparent than in oceanic water under an OA scenario, due to high biomass and sufficient nutrients.With a larger diurnal pH variation range in the future ocean, T. oceanica would be affected more than observed in the present study.Thus, based on our results, the competitive disadvantage for organisms like T. oceanica under fluctuating carbonate chemistry conditions would be amplified in the OA scenario.
Given the decreased growth and elemental production rates of T. oceanica under fluctuating seawater carbonate chemistry in the OA scenario, and its limited ability to dissipate excess excitation energy through NPQ under high light (Strzepek and Harrison, 2004), this species is unlikely to be able to acclimate to coastal habitats, where major fluctuations in light and carbonate chemistry will exist, in the future oceans.In contrast, T. weissflogii appears to be insensitive to, or even benefit from, fluctuating carbonate chemistry.This striking contrast of physiological traits in coastal and oceanic diatoms suggests that the ability to cope with fluctuating carbonate chemistry may play a role in influencing the geographic distributions of species under OA conditions.It is possible that this ability, together with the abilities to cope with nutrient (Irwin et al., 2006), light (Lavaud et al., 2007;Lavaud and Lepetit, 2013;Laviale et al., 2015), and predation pressure (Irigoien et al., 2005), will determine the spatial distribution patterns of species in the future oceans.However, phytoplankton are known to exhibit species-specific response to environmental factors (including OA, fluctuating carbonate chemistry, etc.), and thus more studies on the responses of phytoplankton at the species and community levels are needed to predict such broad biogeographic trends.

Data availability
Data in the present study are available at https://issues.pangaea.de/browse/PDI-13534.

Figure 1 .
Figure 1.Measured pH NBS variation over a diel cycle in the four experimental treatments (LCs, closed triangles; LCf, open triangles; HCs, closed circles; HCf, open circles).Here pH values of triplicate cultures in one experimental day are shown.

Figure 2 .
Figure 2. Specific growth rates of T. weissflogii (a) and T. oceanica (b) under steady (closed bars) and fluctuating (open bars) regimes of ambient (LC) and elevated (HC) pCO 2 levels.Values are means ± SD of triplicate cultures.The different letters indicate significant (p< 0.05) differences among treatments.

Figure 3 .
Figure 3. Production rates of chlorophyll a (a, b), POC (c, d), PON (e, f), and BSi (g, h) of T. weissflogii (a, c, e, g) and T. oceanica (b, d, f, h) under steady (closed bars) and fluctuating (open bars) regimes of ambient (LC) and elevated (HC) pCO 2 levels.Values are means ± SD of triplicate cultures.The different letters indicate significant (p< 0.05) differences among treatments.

Figure 4 .
Figure 4. Chlorophyll-normalized net oxygen evolution rates determined at 6 h after the onset of light for T. weissflogii (a) and T. oceanica (b) under steady (closed bars) and fluctuating (open bars) regimes of ambient (LC) and elevated (HC) pCO 2 levels.Oxygen evolution rates per cell of T. weissflogii (c) and T. oceanica (d) of the four treatments determined 0.5, 6, and 11.5 h after the onset of light.Values are means ± SD of triplicate cultures.The different letters indicate significant (p< 0.05) differences among treatments.

Figure 5 .
Figure 5. Dark respiration rates determined at 6 h after the onset of light for T. weissflogii (a) and T. oceanica (b) under steady (closed bars) and fluctuating (open bars) regimes of ambient (LC) and elevated (HC) pCO 2 levels.Values are means ± SD of triplicate cultures.The different letters indicate significant (p< 0.05) differences among treatments.