Limnol. Oceanogr., 44(3), 1999, 683–690

The effect of variable concentrations of dissolved molecular carbon dioxide, [CO2,aq], on C : N : P ratios in marine phytoplankton was studied in batch cultures under high light, nutrient-replete conditions at different irradiance cycles. The elemental composition in six out of seven species tested was affected by variation in [CO2,aq]. Among these species, the magnitude of change in C : N : P was similar over the experimental CO2 range. Differences in both cell size and day length-dependent growth rate had little effect on the critical CO2 concentration below which a further decrease in [CO2,aq] led to large changes in C : N : P ratios. Significant CO2-related changes in elemental ratios were observed at [CO2,aq] , 10 mmol kg21 and correlated with a CO2-dependent decrease in growth rate. At [CO2,aq] typical for ocean surface waters, variation in C : N : P was relatively small under our experimental conditions. No general pattern for CO2-related changes in the elemental composition could be found with regard to the direction of trends. Either an increase or a decrease in C : N and C : P with increasing [CO2,aq] was observed, depending on the species tested. Diurnal variation in C : N and C : P, tested in Skeletonema costatum, was of a similar magnitude as CO2-related variation. In this species, the CO2 effect was superimposed on diurnal variation, indicating that differences in elemental ratios at the end of the photoperiod were not caused by a transient buildup of carbon-rich storage compounds due to a more rapid accumulation of carbohydrates at high CO2 concentrations. If our results obtained under high light, nutrient-replete conditions are representative for natural phytoplankton populations, CO2related changes in plankton stoichiometry are unlikely to have a significant effect on the oceanic carbon cycle. The elemental composition (C : N : P) of phytoplankton is known to depend on both nutrient concentrations (e.g., Perry 1976; Sakshaug and Holm-Hansen 1977; Hecky et al. 1993) and light regime (Goldman 1980, 1986; Laws and Bannister 1980). Cell stoichiometry may also differ between species, independent of the environmental conditions. In spite of this variability, a constant elemental molar ratio of C : N : P 5 106 : 16 : 1 has been reported by Redfield et al. (1963) for marine plankton in the open ocean when large water masses are considered. The elemental composition of plankton is reflected in a remarkably uniform ratio of changes in the concentration of inorganic nutrients with depth in the deep ocean as a result of the remineralization of biogenic sinking particles (Redfield 1934; Redfield et al. 1963; Takahashi et al. 1985; Andersen and Sarmiento 1994). A constant stoichiometry of marine export production, often called the ‘‘Redfield ratio,’’ is thus commonly used in the calculation of carbon fluxes from nutrient concentrations (e.g., Broecker et al. 1985; Six and Maier-Reimer 1996). In a study of the marine diatom S. costatum, Burkhardt and Riebesell (1997) observed changes in the elemental composition in response to variable CO2 concentrations. These results contradict the commonly accepted notion that elemental ratios of marine phytoplankton are unaffected by CO2 availability. The authors argued that changes in the Redfield ratio may be expected upon the currently observed increase in atmospheric pCO2 if dependence of the elemental composition proves to be a general phenomenon in marine phytoplankton. However, while C : P and N : P varied by up to 65% over the experimental CO2 range of 0.5–38 mmol Acknowledgments We thank A. Dauelsberg and K.-U. Richter for technical support. This is AWI Publication 1584. kg21, the largest changes in the elemental composition occurred at CO2 concentrations lower than typically encountered in ocean surface waters. CO2-related changes in C : N were even smaller and pointed in the opposite direction. Burkhardt and Riebesell (1997) emphasized that the study of more species is a prerequisite to evaluate the biogeochemical relevance of CO2-related changes in elemental composition of marine phytoplankton. The controversial debate regarding inorganic carbon acquisition by marine microalgae (e.g., Riebesell et al. 1993; Laws et al. 1995, 1997; Korb et al. 1997; Nimer et al. 1997; Raven 1997; Tortell et al. 1997; Burkhardt et al. unpubl. data) indicates that species may differ in their mechanisms of carbon uptake. If HCO is in3 volved in inorganic carbon acquisition, either through direct uptake by an energy-dependent transport mechanism or through extracellular, catalyzed conversion of HCO to CO2 3 by carbonic anhydrase (CA), we may expect little sensitivity of the elemental composition to variable [CO2,aq] in the bulk medium. By contrast, irrespective of the inorganic carbon source (CO2 or HCO ) and the mechanism of carbon acqui3 sition (passive diffusion or active transport), cellular carbon demand varies with both growth rate and size of an algal cell. If CO2 is taken up, we may thus expect greater sensitivity of cell stoichiometry to changes in [CO2,aq] in larger species during growth under high light, nutrient-replete conditions. In this study, we tested the effect of CO2 availability on C : N : P ratios in seven species of marine phytoplankton, which covered a wide range of cell sizes. Our main goal was the quantification of changes in elemental ratios in response to variable CO2 concentrations rather than the identification of the underlying mechanisms. In addition, we investigated two other aspects that may be critical to the interpretation of CO2-related effects on the elemental ratios. 684 Burkhardt et al. First, even under light-saturated and nutrient-replete conditions, the rate of photosynthetic carbon assimilation may vary by a factor of .2 within a species in response to variable day length (Burkhardt et al. unpubl. data). As a consequence of enhanced carbon flux into the cell, we expect an increase in the critical concentration below which passive diffusion of CO2 becomes insufficient to satisfy cellular carbon demand (Riebesell et al. 1993; Wolf-Gladrow and Riebesell 1997). Differences in growth rate may thus affect the C : N : P vs. [CO2,aq] relationship. To test this, we incubated each of the species under continuous light and in a 16 : 8 or 18 : 6 light : dark (LD) cycle at six CO2 concentrations. Second, C : N : P ratios in phytoplankton exhibit diurnal variation when grown in an LD cycle as the result of an uncoupling of photosynthetic carbon fixation from nutrient assimilation (Cosper 1982; Cuhel et al. 1984). In experiments of Burkhardt and Riebesell (1997), samples were taken at the end of an 18-h photoperiod. Differences in elemental ratios over the range of [CO2,aq] could thus reflect differences in the ability of a cell to accumulate carbohydrates during the day for subsequent use as an energy reserve in dark respiration. In this case, we might expect no differences in C : N : P at the end of the night between cells growing at variable [CO2,aq]. To test this, we monitored diurnal variation in C : N : P ratios of the diatom S. costatum at two CO2 concentrations during growth at a 16 : 8 irradiance cycle and compared the results to growth under continuous light conditions. Materials and methods The diatoms Phaeodactylum tricornutum, S. costatum, Asterionella glacialis, Thalassiosira weissflogii, Thalassiosira punctigera, and Coscinodiscus wailesii and the dinoflagellate Scrippsiella trochoidea were obtained from the culture collection at the Alfred Wegener Institute. Prior to the experiments, all species were grown under conditions identical to the respective treatments for at least nine cell divisions to ensure adaptation of the cells to CO2 and light supply. Experimental setup, sampling protocol, and analytical methods were the same as in previous studies and are described in detail by Burkhardt and Riebesell (1997). In the following, we will briefly summarize experimental design and measurements. All experiments were performed in 2.4-liter dilute batch cultures at 158C and an incident photon flux density of 150 mmol photons m22 s21. In addition to experiments under continuous light, cells were grown in an 18 : 6 (T. weissflogii) or a 16 : 8 (all other species) LD cycle. Aged, 0.2-mm–filtered seawater from the North Sea was enriched with a trace metal mix and vitamins at concentrations of f/2 medium (Guillard and Ryther 1962). Nitrate, silicate, and phosphate were added to final concentrations of ca. 100, 100, and 6.25 mmol kg21, respectively. Salinity varied between 30.5 and 31.5 psu, depending on the batch of seawater. CO2 concentrations were adjusted by the addition of HCl or NaOH, which resulted in changes in pH and alkalinity at a constant concentration of dissolved inorganic carbon (DIC). To cover an experimental [CO2,aq] range from 1.5 to 37.7 mmol kg21, pH varied from 9.1 to 7.8. It is important to note that the degree of pH variation in our experiments was identical with pH changes that would accompany corresponding changes in [CO2,aq] in the ocean. In ocean surface waters, [CO2,aq] is typically in the range of 10–20 mmol kg21. The incubation bottles were inoculated at low cell densities to permit nine or more cell divisions prior to sampling. During the experiments, ,2% of DIC was taken up by the cells, so pH and carbon speciation was largely unaffected by algal growth. pH was measured with a microprocessor pH meter (WTW, pH 3000) using a combined AgCl/KCl electrode, calibrated with National Bureau of Standards (NBS) buffer solutions. Total alkalinity was determined from linear Gran plots (Gran 1952) after potentiometric titration in duplicate in a temperature-controlled automated system (Metrohm pH-713, coupled with Metrohm Dosimat 665). DIC was measured in duplicate by coulometric titration (UIC, model 5012) in an automated gas extraction system (Johnson et al. 1993). [CO2,aq] was calculated from DIC, total alkalinity, temperature, salinity, and concentrations of phosphate and silicate, assuming dissociation constants according to Mehrbach et