Introduction

By the end of the century, the diffusion of CO2 from fossil fuel combustion into surface oceans is expected to cause CO2 and HCO3 − levels to increase 260% and 20%, respectively1. Beyond the combustion of fossil fuels, upwelling, and riverine discharge, another prominent CO2 source in coastal ecosystems is eutrophication-enhanced microbial respiration2,3,4. The degradation of excessive organic matter can lead to the seasonal accumulation of respiratory CO2 which lowers seawater pH and increases pCO2 to levels not expected in the open ocean until next century (>1,000 µatm4). Shifts in the concentrations of various inorganic carbon sources within the total dissolved inorganic carbon (DIC) pool are likely to elicit a variety of responses from marine flora and fauna. Decreased availability of CO3 2− can inhibit the growth of calcifying organisms5,6,7, while increased availability of CO2 in bulk seawater may benefit some, but not all, photosynthetic organisms8,9,10,11. The photosynthetic organisms most likely to benefit from an increase in CO2 levels are non-calcifying autotrophs whose inorganic carbon uptake is not substrate-saturated at present CO2 concentrations9, or autotrophs which may gain energetic benefit from the downregulation of processes involved in the actively concentrating carbon internally12.

Numerous non-calcified marine autotrophs have been shown to benefit from anthropogenically-induced changes in carbonate chemistry. Marine photosynthetic organisms acquire carbon through the active transport of CO2 and HCO3 − as well as the diffusive uptake of CO2 13. Since HCO3 − is more abundant than CO2 in seawater, many marine autotrophs rely on carbon concentrating mechanisms (CCM) and intracellular or extracellular carbonic anhydrase (CA) to convert HCO3 − to CO2 for use by RuBisCO9, 13,14,15. For marine macroalgae, a variety of chlorophytes, phaeophytes, and rhodophytes are able to utilize HCO3 − and CO2 for photosynthesis14. When exposed to elevated CO2, some chlorophytes such as Ulva rigida and U. lactuca experience increased growth11, 16, 17, while others do not18. Non-calcifying rhodophytes such as Gracilaria lemaneiformis, G. tikvahiae, Chondrus crispus 11, 19, 20, and phaeophytes such as the giant kelp (Macrocystis pyrifera 12) have been shown to benefit from elevated CO2 concentrations. Elevated CO2 can also accelerate the growth of individual species of plankton within multiple classes, including dinoflagellates (Alexandrium fundyense 10, Karlodinium veneficum 21 , Alexandrium ostenfeldii 22), diatoms (Skeletonema costatum 23, Pseudo-nitzschia multiseries 24 , Pseudo-nitzschia fraudulenta 25), and raphidophytes (Heterosigma akashiwo 26). However, not all species within these groups benefit, as is the case of several dinoflagellates10, 22, 26. Additionally, some studies have found that natural plankton community growth and composition will be unaffected by increases in pCO2 to levels predicted by 210027, 28.

The community structure of marine autotrophs is strongly shaped by competition, which can be affected by relative abundance of resources such as nutrients, light, and inorganic carbon. For example, as nutrient loading increases, macroalgae gain a competitive advantage over seagrass29. A similar trend can be found within macroalgal communities, as increased nitrogen loading can favor fast-growing species, such as Ulva spp. over slower-growing ones29 due to the former possessing higher rates of maximum nutrient uptake30. Continued nitrogen loading, however, can shift the competitive advantage in favor of phytoplankton, which often have a higher Vmax, a lower Km, and a higher α than macroalgae31, thus allowing for faster nutrient acquisition and dominance under conditions of extreme nutrient loading rates and extended residence times29. Shifts in the concentration and speciation of inorganic carbon in estuaries may also drive competition among autotrophs. In the presence of high CO2, some species of macroalgae may down-regulate their CCMs, thus permitting more energy to be available for other processes such as vegetative growth9, 11 or may shift towards diffusive uptake of CO2 over use of a CCM to relieve carbon limitation11, 32. Some algal species rely strictly on the diffusive uptake of CO2 or the active transport of HCO3 −, with most species being capable of using both forms of carbon14. Thus, the physiological responses of individual algae to increased CO2 may alter community structure33, 34.

Recently, we have demonstrated that populations of Ulva rigida and Gracilaria tikvahiae from Northwest Atlantic coastal waters experience accelerated growth and likely CO2 uptake when exposed to elevated pCO2 11. The objective of this study was to assess how elevated concentrations of CO2 influences competition among estuarine autotrophs including Ulva rigida, Gracilaria tikvahiae, diatoms, and dinoflagellates. Each macroalgal population was grown with and without elevated levels of pCO2 as well as with and without the other alga, and with and without ambient plankton populations. The growth responses, δ13C signatures, and elemental composition of algae were evaluated at the start and end of experiments performed through the growing season of these macroalgal populations.

Methods

Macroalgae Collection and Preparation

Macroalgae used for this study were collected from Shinnecock Bay, NY, USA (Fig. 1; 40.85°N, 72.50°) during low tide. Permission to access the water and collect the water and macroalgae was received from the Southampton Town Trustees, Southampton, NY, USA, who hold jurisdiction over Shinnecock Bay. Large, well-pigmented, robust fronds of Ulva and Gracilaria were collected and transported to the Stony Brook Southampton Marine Science Center in seawater-filled containers within 15 minutes of collection. Prior research has used DNA sequencing and microscopy to determine that Ulva rigida and Gracilaria tikvahiae are the species of Ulva and Gracilaria present at the same sampling sites used here during summer and fall11. The visual and microscopic analyses during this study affirmed that identification. Due to the plastic nature of macroalgal taxonomic nomenclature as well as the high similarity of ITS sequences among Ulva species35, 36, for the purposes of this study and consistency with prior studies11, we refer to these algae simply as Ulva and Gracilaria. Individual thalli of Gracilaria approximately 5 cm in length were cut from the main plant and placed in a salad spinner to remove debris and epiphytes. Samples were extensively rinsed with filtered (0.2 µm) seawater and placed back into the salad spinner to further remove debris, epiphytes, and excess seawater. Circular sections of similar length of Ulva were cut from large thalli with care taken to avoid the outer, potentially reproductive region of the plant37. Samples of Ulva were prepared using the same cleaning procedures as Gracilaria. All samples were weighed on an A&D EJ300 digital balance (±0.01 g) to obtain initial wet weight in grams. To prevent desiccation, all samples were kept in 100 mL filtered (0.2 µm) seawater-filled containers after spinning and weighing but prior to use in experiments.

Figure 1
figure 1

Map of Shinnecock Bay, NY, USA. The asterisk represents the shallow-water region where macroalgal collections occurred and in situ experiments were performed. All maps were generated using ArcMap 10.4.1 (Esri).

Full size image

In situ Growth Experiments

To assess growth rates of Gracilaria and Ulva within the region of Shinnecock Bay from which they were collected, in situ growth experiments were performed monthly from June through October. Quadruplet, 0.25 m2 incubation cages constructed from 1 cm2 wire mesh were attached to a four-armed (25 cm) umbrella fishing apparatus on a line with surface flotation and a bottom weight to keep the cages suspended at 0.2 m11, 37. Continuous measurements of light and temperature were made using HOBO pendant temperature and light loggers. Thalli of each species of macroalgae were placed in each quadruplet cage for approximately one week in parallel with laboratory experiments (described below) after which thalli were recovered, brought to the lab, and rinsed, spun, re-rinsed, re-spun, and weighed as described above. Samples of Gracilaria and Ulva were frozen for further tissue analysis. Weight-based growth rates for both species were determined using the relative growth rate formula (growth d−1) = (ln Wfinal − ln Winitial)/(Δt), where Wfinal and Winitial are the final and initial weights in grams and Δt is the number of days of the experiment.

Assessing the Effects of Elevated pCO2 and Competition

Five laboratory experiments were performed to assess the effects of competition and elevated pCO2 on the growth of Gracilaria, Ulva, and natural plankton communities during early July, late July, August, September and October. Polycarbonate bottles (2.5 L) were acid washed (10% HCl) and liberally rinsed with deionized water before use. Experimental bottles were placed in an environmental control chamber set to the approximate temperature (~16–21 °C) and light intensity (~400 µmol s−1 m−2) and duration (14 h: 10 h light: dark cycle) present during in situ experiments. Bottles were filled with filtered (0.2µm polysulfone filter capsule, Pall) with the plankton community removed or unfiltered seawater with the full plankton community. For the early and late July, and August experiments, bottles were randomly assigned and dispersed, in triplicate, to one of four treatments: a control with ambient levels of pCO2 (~400 µatm) in filtered seawater (no plankton), a treatment with ambient pCO2 in unfiltered seawater (with plankton), a treatment with elevated pCO2 (~2,500 µatm) in filtered seawater (no plankton), and a treatment with elevated pCO2 in unfiltered seawater (with plankton). Three sets of these bottles were established: One for Ulva, one for Gracilaria, and one with both Ulva and Gracilaria resulting in a total of 36 experimental bottles. For the September and October experiments, bottles were randomly assigned and dispersed to the aforementioned treatments, but in quadruplicate. Additionally, eight bottles were filled with seawater only with four bottles being subjected to ambient pCO2, and the other four being subjected to elevated pCO2. All bottles for each experiment received nutrient additions (50µM nitrate, 3 µM phosphate) at the beginning of the experiment to ensure nutrient replete growth. The nutrient and pCO2 concentrations used during experiments were higher than what is present at the collection site, but are within the range of concentrations present in eutrophic US East Coast estuaries4, 37 and used during prior experiments with Ulva and Gracilaria from Shinnecock Bay, NY, USA11.

Each bottle was aerated via 3.8 × 1.3 cm air diffusers (Pentair) connected to a 1 mL, polystyrene serological pipette inserted to the bottom of each bottle and connected via tygon tubing to an air source. Bottles were subjected to the control (~400 µatm) and elevated (~2500 µatm) levels of pCO2 via a gas proportionator system (Cole Parmer® Flowmeter system, multitube frame) that mixed ambient air with 5% CO2 gas (δ13C = −28‰)5. The gas mixtures were delivered at a net flow rate of 2500 ± 5 mL min−1 through an 18- or 14-way gang valve into the serological pipettes that fit through an opening in the closed cap of the bottle. The delivery rate of gases turned over the volume of the experimental bottles >1,000 times daily5 and bottles were left uncapped but covered with aluminum foil to permit gas exchange Bubbling began two days prior to beginning each experiment allowing pCO2 concentrations and pH levels to reach a state of equilibrium. Experiments persisted for one week. Measurements of pH within bottles were made daily through use of an Orion Star A321 Plus electrode (±0.001) calibrated prior to use with National Institute of Standards and Technology (NIST) traceable standards. DIC concentrations in bottles were measured using an EGM-4 Environmental Gas Analyzer (PP Systems) system that quantifies DIC levels after separating the gas phase from seawater by acidification using a Liqui-Cel Membrane (Membrana)5. As a quality assurance measure, the levels of DIC and pH with Dr. Andrew Dickson’s (University of California, San Diego, Scripps Institution of Oceanography) certified reference material (Batches 142, 147, 151 = 2038, 2014, and 2033 µmol DIC kg seawater−1, respectively) were measured during analyses of every set of samples. The analysis of samples continued only after complete recovery of the certified reference material was attained. The measured values were 104 ± 3.9% of the certified values. Levels of pCO2 (mean of t = initial and t = final, Table 1) were calculated using measured levels of DIC, pH (NIST), temperature, and salinity, as well as the first and second dissociation constants of carbonic acid in seawater38 using the program CO2SYS (http://cdiac.ornl.gov/ftp/co2sys/). The targeted levels of pCO2 resulted in actual pCO2 and pH values of ~400 µatm and ~8.0, respectively, for ambient conditions and ~2600 µatm and ~7.2, respectively, for the elevated CO2 conditions, mimicking the range found seasonally in estuarine environments3, 4, 39.

Table 1 Values of pH (NBS scale), temperature (°C), salinity (g kg−1), pCO2 (µatm), DIC (µmol kgSW−1), HCO3 − (µmol kgSW−1) for Gracilaria and Ulva for June through October experiments.
Full size table

Experiments began with the introduction of macroalgae and nutrients into experimental bottles. HOBO pendant temperature and light data loggers were used to continuously monitor light levels. At the end of experiments, final pH, temperature, and salinity measurements were made and a final DIC was collected and analyzed as described above. After measuring DIC, all macroalgae samples were removed from their respective bottles and rinsed, spun, re-rinsed, re-spun, and weighed as described above. Gracilaria and Ulva samples were placed into small freezer bags for further analyses. Weight-based growth rates for both species were determined as described above. Significant differences in growth rates were assessed using three-way ANOVA with SigmaPlot 11.0, where the main treatments were pCO2 treatment (ambient or elevated), the presence of plankton (filtered or unfiltered seawater), and competition (each macroalgal species alone or in the same bottle). Additionally, one-way ANOVA were used to compare the growth rates of the control group and the in situ experiments.

The growth and composition of the plankton community was assessed during the September and October experiments by removing 50 mL aliquots of seawater from experimental bottles in unfiltered seawater treatments at the beginning and at the conclusion of each experiment and preserving samples with Lugol’s iodine. Aliquots were placed in Sedgewick-Rafter chambers and enumerated using a light microscope, an approach that permitted the quantification of plankton >10 µm37. More than 200 cells were quantified per sample. For the purposes of this study, the most abundant phytoplankton groups were quantified, specifically diatoms and dinoflagellates. Significant differences in abundance were assessed using three-way ANOVA with SigmaPlot 11 where the main treatments were pCO2 (ambient or elevated), Ulva (with or without Ulva), and Gracilaria (with or without Gracilaria).

Tissue Analyses

For carbon (C), nitrogen (N), and stable carbon isotope (δ13C) analyses, frozen samples of Gracilaria and Ulva were dried at 55 °C for 48 h and then homogenized into a fine powder using a mortar and pestle. The total tissue C, N, and δ13C were analyzed using an elemental analyzer interfaced to a Europa 20–20 isotope ratio mass spectrometer at the UC Davis Stable Isotope Facility. Significant differences in tissue content for each species of algae and class of phytoplankton during experiments were assessed using three-way ANOVA within SigmaPlot 11.0 where the main treatment effects were pCO2 treatment (ambient or elevated), the presence of plankton (filtered or unfiltered seawater), and competition (each macroalgal species alone or in the same bottle).

Lastly, we made use of an isotopic mixing model to estimate the use of CO2 and HCO3 − during experiments11. This model considered the δ13C and biomass of macroalgal tissue before and after experiments, the δ13C of the 5% CO2 gas used for the experiments (−28‰), the δ13C of the marine CO2 and HCO3 − pool (−10‰ and 0‰, respectively40,41,42), C fractionation during macroalgal uptake of CO2 and HCO3 − (−20‰ and −10‰, respectively40,41,42), C fractionation during conversion of the 5% CO2 gas bubbled into the experimental containers to HCO3 − (+10‰)40,41,42, and the DIC concentration with and without the addition of the 5% CO2 gas. The latter provides indication of the fraction of DIC contributed by the tanked CO2 gas compared to ambient air. The model assumed that the tanked CO2 reached equilibrium with the total DIC pool, allowing the HCO3 − pool to assume a lighter δ13C signature proportional to the fraction of the DIC pool comprised of tanked CO2 compared to ambient air, an assumption supported by the high turnover rate of seawater by the bubbled CO2 mixture (1000-times daily). Due to the macroalgal tissue being dried and homogenized, it was assumed that the δ13C signature of the macroalgal tissue was representative of the fraction of original tissue with its original δ13C and the tissue grown during the experiment taking on a δ13C signature representative of the DIC pool with a value made proportionally more negative (lighter) by the tanked CO2 gas11. Finally, two sets of mixing models were run for each macroalgal species that estimated their δ13C signature based on exclusively CO2 and exclusively HCO3 − during the experiments11. A one-way ANOVA was used to assess the differences between the measured δ13C signatures of the macroalgae and signatures calculated based on exclusive use of either CO2 or HCO3 −, with Tukey tests used to assess the differences between individual groups.

Results

Gracilaria

The in situ growth of Gracilaria in Shinnecock Bay was found to be similar to and not significantly different from growth rates within the control groups of experiments, with the exception of the early July and August experiment, when experimental growth rates were slightly lower and higher, respectively, than those in situ (One-way ANOVA; p < 0.05; Fig. 2; Supplementary Table S2). The growth rates of Gracilaria within the experimental groups were found to be sensitive to changes in CO2 concentrations (Fig. 2). During experiments in late July, August, and October, the growth of Gracilaria increased significantly when exposed to elevated CO2 concentrations (Three-way ANOVA; p < 0.05; Fig. 2; Supplementary Table S2). On average, growth rates under elevated CO2 were 37% higher and 30% higher than growth under ambient conditions in experimental bottles filled with filtered and unfiltered seawater, respectively (Fig. 2). Growth rates of Gracilaria were not affected by the presence of Ulva and were mostly unaffected by the presence of plankton with the exception of the early July experiment when plankton significantly slowed the growth of Gracilaria (Three-way ANOVA; p < 0.05; Fig. 2; Supplementary Table S2). During the August experiment, there was an interaction between CO2, competition with Ulva, and competition with plankton, whereby elevated CO2 significantly enhanced growth rates within filtered treatments (Three-way ANOVA; p < 0.05; Fig. 2; Supplementary Table S2) but not within unfiltered treatments (Three-way ANOVA; p > 0.05; Fig. 2; Supplementary Table S2). Additionally, in this same experiment, growth was significantly higher under elevated CO2 in treatments without Ulva (Three-way ANOVA; p < 0.05; Fig. 2; Supplementary Table S2), but not in treatments with competition from Ulva, demonstrating that Ulva altered the response of Gracilaria to CO2 in this experiment.

Figure 2
figure 2

Growth rates of Gracilaria exposed to ambient and elevated CO2 conditions, with and without competition from Ulva, and with and without competition from plankton for experiments performed July through October. For three-way ANOVA, CO2 was a main treatment effect during the late July, August, and October experiments. The presence of plankton was a main treatment effect during the early July experiment (see Supplementary Table S2).

Full size image

The δ13C content of Gracilaria was significantly reduced by elevated CO2 delivery, with the average of the ambient and elevated CO2 treatments being, −13‰ and −24‰, respectively (Three-way ANOVA; p < 0.001; Fig. 3; Supplementary Tables S2-S3). Overall, there was no significant difference in δ13C between filtered and unfiltered seawater treatments, regardless of CO2 concentration (Three-way ANOVA; p > 0.05; Supplementary Tables S2-S3). Additionally, there was no significant difference in δ13C caused by exposure to Ulva. Isotopic mixing models demonstrated that, when exposed to elevated CO2 concentrations, the δ13C signatures of Gracilaria (−24‰) were significantly lower than values expected if their C was obtained exclusively from use of HCO3 − (−14‰; Tukey test; p < 0.05; Supplementary Fig. S1), but not significant different than expected from exclusive use of CO2 (−28‰; Tukey test; p > 0.05; Supplementary Fig. S1). On average, the tissue C content of Gracilaria was largely unaffected by CO2 concentration, competition with Ulva, and competition with plankton (Three-way ANOVA; p > 0.05; Fig. 4; Supplementary Tables S2 and S4). However, elevated CO2 was found to have significantly increased the tissue C content relative to the ambient concentration for the late July experiment (Three-way ANOVA; p < 0.05; Supplementary Table S2). Competition with Ulva significantly reduced tissue N of Gracilaria for the August, September, and October experiments, while competition with plankton significantly decreased tissue N for all experiments with the exception of the August experiment (Three-way ANOVA; p < 0.05; Supplementary Tables S2 and S4). Elevated CO2 treatments resulted in decreased tissue N for only the September experiment (S2 and S4 Tables). The tissue C:N ratio of Gracilaria was unaffected by elevated CO2 concentrations (Three-way ANOVA; p > 0.05; Supplementary Tables S2 and S4), but was found to be significantly higher during competition with Ulva during the August experiment and during competition plankton assemblages during the early and late July experiments (Three-way ANOVA; p < 0.05; Fig. 4; Supplementary Tables S2 and S4).

Figure 3
figure 3

δ13C content of Gracilaria and Ulva exposed to ambient and elevated CO2 conditions, with and without competition from Ulva, and with and without competition from plankton for experiments performed July through October. For three-way ANOVA, CO2 was a main treatment effect, on average (see Supplementary Table S2).

Full size image
Figure 4
figure 4

Tissue nitrogen, carbon, and C:N content of Gracilaria exposed to ambient and elevated CO2 conditions, with and without competition from Ulva, and with and without competition from plankton for experiments performed July through October. For tissue N and C:N ratio, CO2, the presence of plankton (un/filtered), and the presence of Ulva were main treatment effects of three-way ANOVA, on average (see Supplementary Table S2).

Full size image
Figure 5
figure 5

Growth rates of Ulva exposed to ambient and elevated CO2 conditions, with and without competition from Gracilaria, and with and without competition from plankton for experiments performed July through October. For three-way ANOVA, CO2 was a main treatment effect during the early and late July, August, and September experiments. The presence of plankton was a main treatment effect during the early and late July experiments. The presence of Gracilaria was a main treatment effect during the early July and August experiments (see Supplementary Table S2).

Full size image

Ulva

The growth rates of Ulva during in situ experiments did not differ statistically from those found within the control treatment of experiments (One-way ANOVA; p > 0.05; Fig. 5; Supplementary Table S2). The response of Ulva to the different variables within the experimental bottles was more complex compared to Gracilaria. Overall, growth by Ulva was significantly higher under elevated pCO2 concentrations and significantly higher in treatments without Gracilaria and competing plankton (Three-way ANOVA; p < 0.05; Fig. 5; Supplementary Table S2). During four of the five experiments (early and late July, August, and September), the growth of Ulva increased significantly when exposed to elevated pCO2 concentration, increasing, on average, 38% and 44% relative to ambient treatments in filtered and unfiltered treatments, respectively (Three-way ANOVA; p < 0.05; Fig. 5; Supplementary Table S2). On average, Ulva growth rates were ~20% lower when grown in the presence of plankton, and 12% lower when grown in the presence of Gracilaria (Fig. 5). During the early July experiment, the presence of plankton depressed the growth of Ulva as did the presence of Gracilaria (Three-way ANOVA; p < 0.05; Fig. 5; Supplementary Table S2). Ulva growth in the presence of plankton was also significantly reduced during the late July experiment (Three-way ANOVA; p < 0.05; Fig. 5; Supplementary Table S2). As independent variables, plankton and Gracilaria did not significantly alter Ulva growth rates during the September experiment, but there was a synergistic interaction between elevated pCO2 and the absence of plankton in slowing Ulva growth (Three-way ANOVA; p < 0.05; Fig. 5; Supplementary Table S2). During the October experiment, the growth of Ulva was not affected by any treatment.

Figure 6
figure 6

Tissue nitrogen, carbon, and C:N content of Ulva exposed to ambient and elevated CO2 conditions, with and without competition from Gracilaria, and with and without competition from plankton for experiments performed July through October. For tissue N and C:N ratio the presence of plankton was a main treatment effect of three-way ANOVA, on average (see Supplementary Table S2).

Full size image

The δ13C content of Ulva was significantly reduced by exposure to elevated CO2 concentrations, with the average δ13C of the ambient and elevated CO2 treatments being −12‰ and −33‰, respectively (Three-way ANOVA; p < 0.001; Fig. 3; Supplementary Tables S2-S3). For the entire study, the δ13C of Ulva was not significantly altered by the presence of Gracilaria or plankton (Three-way ANOVA; p > 0.05; Supplementary Tables S2-S3). The δ13C was, however, found to be significantly lower in treatments with plankton present for the August and September experiments (Three-way ANOVA; p < 0.05; Supplementary Tables S2-S3). Isotopic mixing models demonstrated that when exposed to elevated CO2 concentrations, Ulva δ13C signatures (−33‰) were significantly lower than values expected from exclusive use of HCO3 − (−14‰; Tukey test; p < 0.05; Supplementary Fig. S1) and significantly higher than expected from exclusive use of CO2 (−45‰; Tukey test; p < 0.05; Supplementary Fig. S1). Tissue C content of Ulva was not significantly affected by elevated CO2 concentrations, competition with Gracilaria, or competition with plankton (Three-way ANOVA; p < 0.05; Fig. 6; Supplementary Tables S2 and S4). In contrast, during each experiment tissue N content was significantly lower when Ulva was grown in the presence of plankton, with the exception of the October experiment (Three-way ANOVA; p < 0.05; Fig. 6; Supplementary Tables S2 and S4). The tissue C:N ratio of Ulva was significantly higher in the presence of plankton during each experiment except October (Three-way ANOVA; p < 0.05; Fig. 6; Supplementary Tables S2 and S4).

Figure 7
figure 7

Growth rates of diatoms and dinoflagellates exposed to ambient and elevated CO2 conditions, with and without competition from Gracilaria and/or Ulva. On average, the presence of Ulva and Gracilaria were main treatment effects of three-way ANOVA for diatoms and dinoflagellates, respectively (see Supplementary Table S2).

Full size image

Phytoplankton

Regarding phytoplankton communities, at the onset of the September and October experiments, the dominant phytoplankton >10 µm were diatoms, whereas at the end of experiments, the abundance of diatoms diminished and dinoflagellates became more prominent. The growth rates of diatoms and dinoflagellates were found to significantly decrease and increase, respectively, during exposure to elevated CO2 during the September and October experiments (Three-way ANOVA; p < 0.05; Fig. 7; Supplementary Table S2). Diatoms and dinoflagellate growth rates were also affected by the species of macroalgae present. Diatom growth rates were significantly higher in treatments containing Ulva compared to treatments without (Three-way ANOVA; p < 0.05; Fig. 7; Supplementary Table S2). Dinoflagellates growth was significantly decreased in the presence of Gracilaria (Three-way ANOVA; p < 0.05; Fig. 7; Supplementary Table S2).

Figure 8
figure 8

Responses and interactions of various estuarine primary producers to eutrophication, ocean acidification, and competition under three scenarios: (a) Nutrient loading only, with competition. (b) Low to elevated CO2 and nutrient loading, without competition. (c) Low to elevated CO2 and nutrient loading, with competition.

Full size image

Discussion

During this study, elevated CO2 concentrations significantly enhanced the growth rates of Gracilaria, Ulva, and dinoflagellates, but not diatoms. For Gracilaria, growth rates were largely unaffected by the presence of Ulva and plankton whereas the growth rates of Ulva were significantly depressed when grown with Gracilaria or the full plankton community. Among the phytoplankton, diatom growth benefited from the presence of Ulva, while the growth rates of dinoflagellates were slowed by Gracilaria. For both macroalgae, tissue δ13C was significantly lowered by elevated pCO2 while tissue N content was reduced by competition with the other macroalgae species and/or plankton. While these experiments were performed within bottles, the rapid turnover of the dissolved gas pools yielded growth rates of macroalgae that were nearly identical to parallel thalli concurrently measured in an ecosystem setting evidencing the realistic nature of conditions during experiments. Collectively, these findings provide novel insight regarding the outcomes of competition among primary producers under high CO2 conditions.

Most macroalgae are capable of active transport of HCO3 − or CO2 into their CCM or the diffusive uptake of CO2 13. High CO2 concentrations may cause macroalgae to down-regulate CCMs that convert HCO3 − to CO2 16, 19, 43, 44 resulting in more energy available for other processes such as vegetative growth9, 11. The amount of energy saved by this process is not fully clear, as the process depends on several external factors, such as PAR, and internal factors, such as type of CCM used by the macroalgae, or the potential leakage of carbon dioxide from the CCM45. The δ13C signatures of macroalgae during this study suggested these species switched from mostly HCO3 − to more CO2 use and potentially downregulated their CCMs. Values prior to the start of the experiments (−12–13‰) were reflective of HCO3 − and CCM use whereas the more negative values of macroalgae at the end of the experiment (−23.6 ± 5‰ and −33.5 ± 5‰ for Gracilaria and Ulva, respectively) were within the range expected of macroalgae relying more on the diffusive uptake of CO2 12, 40, 46 using isotope mixing models to account for the lighter CO2 gas used in experiments11. It is also possible that higher pCO2 alleviated inorganic C limitation and enhanced growth rates. Mercado et al.32 reported that U. rigida and U. compressa (formerly Enteromorpha) do not receive enough CO2 through diffusive uptake at current CO2 levels, a finding consistent with the enhanced growth of Ulva during this study and supported by the shift in δ13C during this study for both Ulva and Gracilaria. Regardless, the enhanced growth rates for these macroalgae under higher CO2 indicate that inorganic C limitation was alleviated.

Consistent with prior studies of macroalgae, changes in CO2 levels did not alter tissue C and N content11, 47 and competition with other autotrophs did not alter their C content. In contrast, competition with other autotrophs resulted in significantly decreased N content and decreased tissue C:N ratios for Gracilaria and Ulva. Both macroalgal species are able to rapidly assimilate and store nitrate48, 49 and have been shown to experience enhanced tissue N content when exposed to excessive nitrate concentrations50, 51. Compared to Gracilaria, Ulva is capable of undergoing more rapid growth in eutrophic settings29, 37 due to a high maximum rate of uptake of nutrients such as nitrate30. Phytoplankton are superior competitors for N compared to macroalgae29, 31. The significant declines in N content of macroalgae when grown with plankton and elevated C:N ratios of macroalgae at the end of experiments (15–40), despite the high levels of N present at the start of experiments (50 µM), affirms the role of N as a limiting element in this52 and other estuaries53, and suggests this N was likely depleted over the course of the experiment. This is almost certainly the case in experiments with the full plankton community intact as uptake rates of plankton communities can exceed 25 µmol L−1 day−1 in Shinnecock Bay52. The precise outcomes of competition among estuarine autotrophs exposed to high CO2, therefore, will be partly dependent upon ambient nutrient supplies.

Beyond tissue content of macroalgae, the importance of both N and pCO2 in shaping algal community composition was also evident in the competitive growth responses of macroalgae. The ability of macroalgae to respond to shifts in the ratio of HCO3 − to CO2 in seawater may prompt algae capable of using both inorganic carbon species to gain a competitive advantage over algae restricted to only HCO3 −. Cornwall et al.34 found that macroalgal abundance along a CO2 gradient at Vulcano, Italty varied according to the inorganic carbon uptake strategy of the algae. During that study, five macroalgae species capable of using HCO3 − and CO2 increased with abundance as CO2 concentrations increased, as well showed a decline in tissue δ13C associated with increased CO2 use. However, calcifying macroalgae, as well as species incapable of using CO2, decreased with abundance as CO2 concentrations increased34. During the present study, Ulva and Gracilaria, when exposed to elevated CO2, had increased growth and declines in tissue δ13C associated with increased CO2 use, which indicates that both may gain a competitive advantage over species incapable of adjusting their inorganic carbon physiology in response to increases in CO2 in seawater. But besides carbon use strategies, competition for nutrients is also a key factor to consider. Nutrient loading favors fast-growing macroalgae with rapid uptake rates of nutrients over slower-growing counterparts29, 30. The growth rates of Ulva were, on average, three-times faster than Gracilaria during experiments and thus, despite a 55–60% lower tissue N content, had a significantly larger N demand, making Ulva more prone to N limitation, especially when placed in competition with other autotrophs. This hypothesis is supported by the C:N ratios of Ulva which were significantly higher than those of Gracilaria throughout this study (p < 0.001; T-test), suggesting Ulva was more N-limited. Similarly, the presence of plankton, which are able to outcompete macroalgae for nutrients, may have further depleted nutrient concentrations during experiments, thus causing the decreased growth of Ulva in unfiltered treatments for some of the experiments29. Again, this hypothesis is supported by the significant increase in the C:N ratio that Ulva experienced when grown in the presence of plankton communities. Collectively these findings suggest that while pCO2 enhances the growth of Ulva and Gracilaria, the slower-growing Gracilaria is better adapted for persisting at more dynamic nutrient concentrations than Ulva 30, 54. In a field experiment by Fujita (1985)55, when N was introduced in pulses every five days, Gracilaria tikvahiae was able to outcompete Ulva lactuca in mixed macroalgal beds, despite the latter possessing a more rapid N uptake rate. Furthermore, Gracilaria vermiculophylla, normally found in the West Pacific, has invaded northern European estuaries as early as 2002, and has become among the most abundant macroalgae in the region, despite competition with Ulva and other ephemeral algae56. Despite these instances, were nutrients continuously added during experiments of the current study, it is plausible that the growth of Ulva would have been less affected by other autotrophs. Hence, the outcome of competition among estuarine autotrophs exposed to high CO2 depend, at least in part, on ambient N levels.

Dinoflagellates experienced more rapid growth when exposed to high CO2 while diatoms did not. Results from prior studies suggest that the response of plankton communities to elevated CO2 concentrations are likely to depend on the species present but that dinoflagellates are more prone to C-limitation than diatoms as dinoflagellates possess form II RubisCO, which has a low affinity for CO2 57, 58. The dinoflagellates Protoceratium reticulatum 59, Karlodinium veneficum 21 and Karenia brevis 60 all grow more rapidly under high CO2 as do Alexandrium species from Europe (Alexandrium minutum 61; Alexandrium ostenfeldii 22 and the North America (Alexandrium catenella 21, 62; Alexandrium fundyense 10). While the general response of diatoms to elevated CO2 also appears to be species-specific, they seem to be generally less sensitive to changes in pCO2. Dozens of diatom species realize maximal growth rates under a wide range of pH/ pCO2 levels27, 63,64,65, although elevated CO2 enhances the growth rates of some species including Pseudo-nitzschia fraudulenta 25, Pseudo-nitzschia multiseries 24, and Chaetoceros debilis 66. Hence, the finding that CO2-stimulated growth of dinoflagellates but not diatoms are generally consistent with prior studies, but specific responses will depend on, among other factors, nutrient levels, the species of plankton present within a community, as well as competition with other autotrophs. Given dinoflagellates are responsible for most harmful algal blooms (HABs67) and that HABs are common within eutrophic settings68, the findings here suggest that high CO2, eutrophic estuaries may be more likely to host HABs with negative ecosystem consequences10.

Diatom and dinoflagellate growth rates were also affected by macroalgae with dinoflagellates growth being inhibited by Gracilaria but Ulva promoting the growth of diatoms. Prior studies have found that dinoflagellates in temperate estuaries are vulnerable to allelopathic inhibition by macroalgae69, 70 and Gracilaria spp. have been shown to allelopathically depress dinoflagellate growth rates71, 72. While Ulva has been found to allelopathically inhibit the growth of individual dinoflagellate species in culture69, during this study Ulva was found to have no effect on dinoflagellates but promoted the growth of diatoms. This finding indicates that Ulva may generally promote a succession within phytoplankton communities from dinoflagellates to diatoms, potentially via the remineralization of nutrients73 that promotes the growth of diatoms. The growth promotion of diatoms may be associated with the ability of Ulva to release and regenerate nutrients such as ammonium and phosphate73, 74. Another possibility is that vitamin B12-producing epiphytic bacteria on Ulva may have promoted the growth of diatoms. Diatoms are unable to synthesize vitamin B12 and as such, require bacteria for the production of the vitamin75, 76. Udell et al.77 found samples of Ulva lactuca in the same contiguous water body as the study site to be rich in vitamin B12 likely due to epiphytic bacteria. It is possible that the synthesis of vitamin B12 by epiphytic bacteria could have promoted the growth of diatoms in treatments containing Ulva.

There are numerous ecosystem implications of the overgrowth of macroalgae, such as Ulva and Gracilaria, due to the ability to outcompete autotrophs due to increased nutrient loading and CO2 concentrations. The overgrowth of bloom-forming macroalgae has been shown to have negative effects on seagrass meadows29, 78, kelp forests79, coral reefs80, 81 and even phytoplankton communities69, 70, 72. Although seagrass can experience enhanced growth in the presence of elevated CO2 concentrations8, increased nutrient loading favors macroalgal growth that can lead to the demise of seagrass29, 82 to the detriment of invertebrate and fish species that use seagrass for food, cover, and as nurseries82,83,84,85,86. The overgrowth of macroalgae can also directly cause mortality in some invertebrates87,88,89. Aside from the direct deleterious effects of ocean acidification on coral reefs and calcifying invertebrates90, continued eutrophication and ocean acidification may allow fast-growing macroalgae to overgrow substrate used by coral91. Adding to this point, Diaz-Pulido et al.92 found that the highest abundance of macroalgae on inshore reefs were species with CCM, but capable of using HCO3 − and CO2 for photosynthesis. If these algae are capable of increased growth under elevated CO2, it could potentially pose another threat to the future health of inshore reefs. However, offshore reefs may be more vulnerable to macroalgal overgrowth due to high abundances of macroalgae that strictly use CO2, which will directly benefit from elevated CO2 in the near future92. In sum, the benefits experienced by macroalgae as the result of increased CO2 concentrations can directly and indirectly harm a multitude of coastal ecosystems, as well as the organisms that reside within them.

Traditionally, nutrient loading has been considered a key factor structuring the dominance of autotrophs in estuaries, with seagrasses dominating estuaries with lower nutrient loads and phytoplankton outgrowing seagrasses and macroalgae in eutrophic systems with extended residence times29 (Fig. 8a). This and prior studies now allow the co-effect of CO2 to be considered in structuring autotrophic communities in estuaries. Among phytoplankton, dinoflagellates benefited more than diatoms from elevated CO2 during this and prior studies10, 62 (Fig. 8b) and Ulva and Gracilaria grow faster when exposed to elevated levels of CO2 11 (Fig. 8b). When competition is considered, the ability of Ulva and Gracilaria to benefit from high CO2 and to inhibit the growth of competing phytoplankton via allelopathy, may allow macroalgae to dominate high nutrient, high CO2 estuaries69, 70, 72 (Fig. 8c). Under the single large dose of nutrients used in the experiments presented here, Gracilaria was the ultimate ‘winner’ within experimental treatments with high nutrients and high CO2 (Fig. 8c). Factors such as continuous nutrient loading and shading would likely alter the outcomes of competition.