Introduction

Anthropogenic carbon emissions have led to atmospheric carbon dioxide (CO2) rising by 40 % since pre-industrial times (Raven et al. 2005). By the end of this century, atmospheric CO2 is predicted to double from current levels (Meehl et al. 2007; Collins et al. 2013). The rise in oceanic CO2 concentration that follows is projected to decrease seawater pH by 0.3–0.4 units (Caldeira and Wickett 2003; Feely et al. 2004). This reduction can alter the carbonate chemistry of seawater in terms of the relative proportion of the dissolved inorganic carbon (DIC) species. Current concentrations of CO2 and HCO3 − in seawater are 8 and 1650 µmol kg−1 seawater, respectively (Koch et al. 2013). Under the projected decrease in seawater pH, the proportion of CO2 will have greater proportional increase (>250 %) than the other DIC constituents (\({\text{HCO}}_{ 3}^{ - }\): 24 % and \({\text{CO}}_{ 3}^{ 2- }\): −61 %) (Koch et al. 2013). The higher concentration of utilisable carbon for photosynthesis (CO2 and HCO3 −) in acidified seawater may benefit marine macrophytes that are limited by the DIC concentration under current conditions (Beer et al. 2002).

Seagrasses can be carbon-limited at the seawater DIC composition under current CO2 concentrations, given other conditions, such as light, nutrient availability and water temperature are non-limiting (Beer and Koch 1996; Thom 1996; Zimmerman et al. 1997; Invers et al. 2001). Most seagrasses utilise the C3 metabolism for carbon fixation (Koch et al. 2013). Elevated partial pressure of CO2 (pCO2) can increase carboxylation rates while reducing oxygenation rates of ribulose-1,5-bisphosphate carboxylase–oxygenase (Rubisco), the initial carboxylating enzyme in C3 plants (Bowes and Ogren 1972; Koch et al. 2013). Furthermore, the predominant DIC species, \({\text{HCO}}_{ 3}^{ - }\), appears to be less efficiently utilised in seagrasses—the increase in photosynthetic rates was much higher when seagrasses were enriched with CO2 than with \({\text{HCO}}_{ 3}^{ - }\) (Sand-Jensen and Gordon 1984; Durako 1993; Beer and Koch 1996; Invers et al. 2001). Although seagrasses have been shown to possess carbon-concentrating mechanisms (CCMs) to more efficiently utilise \({\text{HCO}}_{ 3}^{ - }\), whether these CCMs could effectively saturate the seagrasses to meet their DIC requirements under natural conditions remains to be seen (Beer et al. 2002; Koch et al. 2013). Overall, it is thought that higher pCO2 not only increases passive diffusion of CO2 for carbon fixation, but also lowers the loss of fixed carbon through photorespiration (Long et al. 2004). Laboratory and mesocosm experiments conducted over the short and medium term have shown an optimisation of photosynthetic performance, such as light requirements, photosynthetic efficiency and pigment content in response to CO2 (Zimmerman et al. 1997; Jiang et al. 2010; Campbell and Fourqurean 2013b). This can result in higher rates of carbon fixation with flow-on effects to growth rate, carbohydrate content, biomass and reproductive output (Zimmerman et al. 1997; Jiang et al. 2010; Campbell and Fourqurean 2013b). In the field, higher seagrass productivity and biomass have been observed near natural CO2 vents, suggesting that acidification of seawater would benefit seagrass meadow productivity over the long term (Hall-Spencer et al. 2008; Fabricius et al. 2011; Russell et al. 2013).

Different seagrass species might vary in the manner and extent to which they respond to CO2 enrichment. No previous studies have directly compared species responses to CO2 enrichment, but responses to CO2 depletion indicate that species are not affected uniformly by changing pCO2 (Invers et al. 1997; Beer et al. 2006). This makes it difficult to determine whether findings are related to species or methodological differences. Most studies had focussed on temperate species, such as Zostera marina (Thom 1996; Zimmerman et al. 1997; Palacios and Zimmerman 2007), Zostera noltii (Alexandre et al. 2012) and Posidonia oceanica (Invers et al. 2002). Amongst temperate species, Invers et al. (2001) demonstrated that pCO2 enhancement of photosynthesis was higher in Pacific species (Z. marina and Phyllospadix torreyi) than in Mediterranean species (P. oceanica and Cymodocea nodosa). The few studies on tropical seagrasses yielded mixed results. For example, Jiang et al. (2010) showed increased growth and productivity in T. hemprichii, while T. testudinum showed little change in biomass and productivity to increased pCO2 (Durako and Sackett 1993; Campbell and Fourqurean 2013a). Hence, differential response to CO2 enrichment might exist between and within multi-species tropical seagrass meadows.

Differences in carbon utilisation and allocation strategies exist amongst tropical seagrass species (Hemminga and Duarte 2000; Uku et al. 2005). Species-specific differences in DIC uptake mechanisms would result in varying abilities amongst species to utilise the extra DIC (Invers et al. 2001; Uku et al. 2005; Campbell and Fourqurean 2013b). Species-specific carbon allocation strategies could affect how responses to CO2 enrichment manifest at the plant scale. For example, in species that invest a greater proportion of biomass to belowground tissue, such as Halodule uninervis and Thalassia hemprichii, there would be a higher metabolic demand on aboveground tissue for photosynthetic carbon fixation (Terrados et al. 1999; Hemminga and Duarte 2000; Tanaka and Nakaoka 2007). Increased availability of CO2 in seawater could allow for increasing photosynthetic capacity (e.g. more chlorophyll pigments, enhanced shoot growth) and/or increased storage of carbohydrates to support respiratory demands (Zimmerman et al. 1997; Jiang et al. 2010). In addition, small-bodied ephemeral species, such as Halodule uninervis, exhibit short turnover of leaves, while bigger and more persistent species such as Cymodocea serrulata and Thalassia hemprichii have longer shoot plastochrone intervals (Hemminga and Duarte 2000). Turnover rates of assimilated carbon could influence carbon demand (Arp 1991; Hemminga and Duarte 2000). Thus, various measures of productivity, such as tissue growth rates, carbohydrates storage or shoot production could vary amongst co-occurring species in response to CO2 over different timescales.

Productivity of seagrass meadows is central to their ecological functions as a food source, including for megafauna such as dugongs and turtles, in bio-sequestration (“blue carbon”), and substrate stabilisation (Duarte and Chiscano 1999; Gacia and Duarte 2001; Fourqurean et al. 2012; Vafeiadou et al. 2013). Understanding how productivity responses to CO2 enrichment vary amongst species is vital for predicting future ecological change. In the present study, we quantified the photosynthetic and growth responses of three tropical seagrass species to increasing pCO2 levels, bracketing the range of different end-of-century emission scenarios as predicted by IPCC (2013). This allows for the quantification of the response to pCO2 levels in seagrass productivity and growth. The three species examined, Halodule uninervis, Cymodocea serrulata and Thalassia hemprichii, are common seagrasses found in the tropical Indo-Pacific region with contrasting growth strategies, ranging from rapid growth in H. uninervis to slow growth in C. serrulata and T. hemprichii (Hemminga and Duarte 2000). It was hypothesised that pCO2 enrichment would increase photosynthetic and growth rates, but rate of responses may vary between species due to varying carbon uptake and allocation strategies (Campbell and Fourqurean 2013b).

Materials and methods

Experimental species

Seagrasses were collected two to four weeks prior to the start of the experiment. Seagrass species Cymodocea serrulata and Halodule uninervis were collected from the intertidal meadow at Cockle Bay, Magnetic Island, Northern Great Barrier Reef (19°10.88′S, 146°50.63′E) in March 2013. Average daily and average maximum photosynthetically active radiations (PAR) at this site are 385 and 961 μmol m−2 s−1, respectively (Collier, unpublished). Intact plugs of H. uninervis and sediment were collected with a trowel and placed into a plastic pot lined with a plastic bag. The bag was pulled up and secured over the seagrass to prevent moisture loss during transport. C. serrulata was collected by excavating intact shoots with connected horizontal rhizomes from the sediment before placing into seawater-filled containers for transport to aquaria. Thalassia hemprichii was collected from Green Island in the Northern Great Barrier Reef (16°45.37′S, 145°58.19′E), using a similar method to C. serrulata. At this site, average daily PAR was 344 μmol m−2 s−1 and average maximum PAR was 841 μmol m−2 s−1, respectively (Collier C, unpublished). Average water temperatures at Cockle Bay (2005–2012) and Green Island (2003–2012) were 26.2 and 26.6 °C, respectively (McKenzie et al. 2014). Seagrasses were planted into orchid pots lined with a pool filter sock, in a mud and sand (roughly 20:80) mixture, within 2 days of collection. For acclimation, all species were kept in an outdoor flow-through aquarium prior to the experiment, under average light levels of 350 μmol m−2 s−1, average seawater temperature of 25 °C and salinity at 35 ppt.

Experimental set-up

Seagrasses were exposed to four different seawater pCO2 concentrations in a flow-through system for two weeks (Table 1). The experiment was conducted in an indoor flow-through aquarium system at the Australian Institute of Marine Sciences, Townsville. Sixteen glass aquaria with four replicates for each treatment (working volume 18 l) were supplied with fresh filtered seawater from four header tanks. Each aquarium contained all three species. Two sub-replicate pots of each species were placed in each aquarium. pH levels in the header tanks were monitored, as a proxy to control for CO2 input, with eight potentiometric sensors (±0.01 pH unit) calibrated on the NBS scale. The sensors are connected to a feedback control system that regulates pH levels via a CO2 gas injection system (AquaMedic, Germany). Pumps and diffusers installed in mixing tanks and experimental aquaria ensured thorough mixing of CO2. Additional pH readings were taken regularly with a hand held pH probe (pH probe: Eutech, USA; console: Oakton, USA) and compared to Tris seawater standards (Batch 10, Supplied by A. Dixon, Scripps Institute of Oceanography). Water temperature remained constant throughout the experiment around 24 °C (Table 1). Water samples, taken every 5 days, were analysed for dissolved inorganic carbon (DIC) and total alkalinity (AT) concentrations using a Vindta 3C analyser. Carbonate system parameters (Table 1) were calculated by measured values of AT, DIC, temperature and salinity by USGS CO2calc software (Robbins et al. 2010). Illumination was provided with LED lamps (Aqua Illumination) mounted about 40 cm above the aquaria, providing 400 μmol m−2 s−1 of light set on a 12-h light/dark photoperiod. Duplicate water samples collected from each individual aquaria every 5 days were filtered (0.45 μm pore size) before they were analysed for dissolved inorganic nitrogen and phosphorus concentration according to Ryle et al. (1981).

Table 1 Measured and calculated parameters, and average nutrient concentrations for control and three enriched pCO2 treatments
Full size table

Photosynthetic response

Photosynthetic rates and respiration of the second youngest leaf (rank 2) of a haphazardly chosen shoot from each pot were measured using optical oxygen sensors (“optode”, PreSens, Sensor spots-Pst3) and a PreSens Oxy 4 four-channel fibre-optic oxygen meter after two weeks. While the authors acknowledge that seagrasses could be sensitive to physical manipulations such as removing leaves (Schwarz et al. 2000), care was taken to reduce the impact on leaves such as using the whole leaf and gently rubbing epiphytes off with fingers instead of scrapping with a blade. Small transparent acrylic chambers (200 mL) were set in an array of four (i.e. four separate chambers allowing four parallel measures) and incubated at 25 °C water temperature using a flow-through water system connected to a water bath (Lauda, Ecoline RE 106). Stirrer bars placed within the chambers provided even stirring. The leaves were held upright in the chamber to mimic natural orientation. Oxygen consumption (dark respiration) was measured over a 20-min period in the dark. Photosynthetic rates were then measured on the same leaf over a series of light steps (10, 30, 70, 110, 220, 400, 510 μmol m−2 s−1) (Aqua Illumination LED), with each light step lasting 20 min. Seawater within the chambers was replaced with fresh media every two to three steps. Oxygen concentration data in the chambers were logged every 5 s, and respiration and production rates were calculated by fitting a linear regression. Rates were normalised to the dry weight of the leaf. Leaves were dried at 60 °C for 48 h before weighing. Initial periods of incubation (~5 min) prior to stabilisation of photosynthetic rates were omitted from regressions. Each optode was calibrated according to Collier et al. (2011).

Net productivity (NP) was taken to be the photosynthetic rate measured at 400 μmol m−2 s−1, which was the experimental light level. Energetic surplus (P G:R) was calculated as the ratio of gross productivity (sum of net photosynthetic rate and dark respiration rate) to dark respiration rate (Zimmerman et al. 1997). To determine photosynthetic parameters, photosynthesis versus irradiance (P–E) data plots were fitted to the adapted hyperbolic tangent model equation of (Jassby and Platt 1976):

$$P = P_{\hbox{max} } \times \tanh \left( {\frac{{\alpha P_{\hbox{max} } }}{E}} \right)$$

where P max is the maximal photosynthetic rate (mg O2 g−1 DW h−1), E is irradiance (μmol m−2 s−1), and α described photosynthetic efficiency via the gradient of the curve at limiting irradiances (mg O2 μmol−1). Saturating irradiance (E k) is the light level at which photosynthesis initially reaches the maximum rate, and compensation irradiance (E c) is the light level when photosynthetic rate is equal to respiration rate.

Determination of growth rates

Growth was measured following Short and Duarte (2001). All shoots from each pot were marked at the top of the sheath with a needle at the start of the experiment. At the end of the experiment, the shoots were harvested. The length of new tissue growth was excised, dried at 60 °C for 48 h and weighed for determination of weight of new leaf growth. Leaf tissue growth was normalised to the aboveground biomass of its respective pot to derive relative leaf growth rates (RGR).

Specific leaf area (SLA) was calculated from biomass and areal measurements of leaves. Specific leaf area refers to the total leaf area normalised by the total biomass of the leaves and could be used to infer whole-plant changes in leaf biomass and area in response to pCO2 enrichment (Chiariello et al. 1989). Leaves were separated from shoots and placed on a flat surface. Areal measurement of leaves was then carried out by capturing a clear image of all the leaves and analysing with CPCe software (version 3.6) (Kohler and Gill 2006). Finally, the leaves were dried at 60 °C for 48 h and weighed to obtain biomass measurements.

Chlorophyll content

A young mature leaf (rank 2) from each pot was collected and stored immediately at −20 °C at the end of the experiment. To determine chlorophyll concentration, a 10- to 15-mm section of leaf was cut from the middle of a fully mature leaf and the width of the leaf segment was measured using a pair of callipers. The leaves were blotted dry and weighed before they were ground in a chilled mortar. Depending on the species and the weight of the leaf segment, 5–6 mL of cold (4 °C) 90 % acetone was added to extract chlorophyll from the sample. The solution was gently shaken, left in the dark to extract for 24 h at 4 °C and then centrifuged at 2680 g for 4 min to settle the pellet. The extract was measured for chlorophyll concentration according to Granger and Izumi (2002).

Non-structural carbohydrate (NSC) content

Roots and rhizomes were dried at 60 °C for 48 h, before being finely ground in a bead beater (Daintree Scientific). Four replicate samples per treatment and species were sent to the Agriculture and Food Sciences laboratory in University of Queensland for non-structural carbohydrate content analysis. Briefly, soluble carbohydrates were extracted twice with 80 % ethanol at 80 °C for 10 min from 200 mg of ground plant material. Extracts were then passed through a de-colourising column to remove phenolic compounds. After acid hydrolysis, the amount of soluble carbohydrates was assayed with ferricyanide reagent and absorbance measured on a UV–Vis spectrophotometer at 420 nm (McCleary and Codd 1991).

Starch content was analysed according to Karkalas (1985). Residue from the soluble carbohydrate extraction was solubilised in boiling water. After cooling to room temperature, samples underwent enzyme digestion where amylase and amyloglucosidase were added. After incubation, the concentration of glucose is measured using a commercially available glucose oxidase/peroxidase (GOPOD) testing reagent (Megazyme). Absorbance was then measured at 510 nm.

Total non-structural carbohydrate (NSC) content, which was the sum of the amount of soluble carbohydrate and starch content, was expressed as milligrams dry weight−1 of tissue.

Statistical analyses

All statistical analyses were carried out with R software (R Development Core Team 2011). Changes in photosynthetic and growth responses were tested using linear models with average pCO2 levels for each treatment as explanatory variable. Data from sub-replicate pots from each tank were averaged for the analysis. Assumptions of homogeneity of variances and normality were checked using box plots and residual plots. To satisfy the assumptions, photosynthetic efficiency (α) and compensation irradiance (E c) for T. hemprichii were square-root-transformed prior to analysis. One data point was identified as an outlier (>2 SD from mean of remaining replicates) in each of the P G:R and α dataset and was subsequently removed. To examine species differences in productivity and growth responses to increasing pCO2, confidence intervals (CI) of the slopes (degree of response per 100 µatm rise in pCO2) from linear models were calculated and compared.

Results

Experimental parameters

Water temperature (23.7–24.0 °C) and salinity (35 ppt) in the experimental tanks were near-constant throughout the experiment (Table 1). Carbonate system parameters of the enriched pCO2 treatments remained well within the target range (control pCO2 = 442 ± 6 μatm; low pCO2 = 694 ± 20 μatm; intermediate pCO2 = 884 ± 52 μatm; high pCO2 = 1204 ± 59 μatm) (Table 1). Inorganic nutrient concentrations were similar between tanks and averaged to an ammonium concentration of 0.22 ± 0.01 μM, nitrate concentration of 0.88 ± 0.3 μM and phosphate concentration of 0.19 ± 0.02 μM.

Photosynthetic performance

Carbon dioxide enrichment increased seagrass net productivity (NP). Under the chosen light level (400 μmol m−2 s−1), NP significantly increased with increasing pCO2 levels for all species (Fig. 1; Table 2). Across species, the increase in NP ranged from 0.757 to 1.040 mg O2 g−1 DW h−1 for every 100 µatm increase in pCO2; however, no species difference in the slope was detected (based on overlapping confidence intervals) (Table 2).

Fig. 1
figure 1

Linear model fits (dotted lines indicate 95 % confidence intervals) for net productivity and energetic surplus (P G:R) of C. serrulata, H. uninervis and T. hemprichii in response to pCO2 enrichment. N = 4

Full size image
Table 2 Linear models for all response variables measured
Full size table

Energetic surplus, or gross photosynthetic to respiration ratios (P G:R), significantly increased with increasing pCO2 for all three species (Fig. 1; Table 2). No distinct differences in the slopes (0.32–0.47 units as pCO2 increased by 100 µatm, Table 2) indicated that P G:R responses in the different species were similar.

Photosynthetic rates in all three species exhibited typical P–E (PAR) response curves. Photosynthetic rates increased linearly (initial slope, α) with light under limiting irradiances, before levelling off at the maximum photosynthetic rate (P max) past saturating irradiance (E k). Photosynthesis–irradiance (P–E) curves demonstrated a good fit (R 2 > 0.85; P < 0.05) to the adapted hyperbolic tangent model.

Increasing pCO2 levels significantly increased maximum photosynthetic rates (P max) for all three species (Table 2; Fig. 2). Maximal photosynthetic rates (P max) increased by 0.677–0.929 mg O2 g−1 DW h−1 for every 100 µatm rise in seawater pCO2. Photosynthetic efficiency (α) significantly increased with pCO2 levels across all species (Table 2; Fig. 2). Photosynthetic efficiency increased by 0.004–0.013 with every 100 µatm rise in pCO2 level across all species.

Fig. 2
figure 2

Parameters derived from P–E curves. Top row—maximal photosynthetic rates (P max); second row—photosynthetic efficiency (α); third row—saturating irradiance (E k); bottom row—compensation irradiance (E c). Data were fitted with linear models (dotted lines 95 % confidence intervals). N = 4

Full size image

Saturating irradiance (E k) was not significantly altered by the pCO2 treatments (Table 2; Fig. 2). Increasing pCO2 enrichment reduced compensation irradiance (E c) for C. serrulata and H. uninervis (Table 2; Fig. 2); however, in T. hemprichii, E c was not affected by pCO2 enrichment (Table 2; Fig. 2).

Overall, most photosynthetic parameters responded significantly to pCO2 increase. Although some variation exists in the slopes, overlapping CIs indicated that species differences were non-significant (Table 2).

Plant-scale responses (leaf growth and rhizome carbohydrates)

Leaf growth responses to pCO2 enrichment differed between species. C. serrulata did not show differences in growth rates with increasing pCO2 levels (Fig. 3; Table 2). By contrast, growth rates increased with pCO2 enrichment for H. uninveris (Fig. 3; Table 2) and T. hemprichii (Fig. 3; Table 2). Slopes for relative leaf growth rates (RGR) were about 0.001 units for every 100 µatm increase in pCO2 for both species.

Fig. 3
figure 3

Linear model fits (dotted lines indicate 95 % confidence intervals) for relative growth rates of C. serrulata, H. uninervis and T. hemprichii in response to pCO2 enrichment (mgDW mg−1DW day−1). N = 4

Full size image

For plant-scale response to pCO2, amongst the three species only T. hemprichii displayed an increase in specific leaf area (SLA; leaf area per unit dry weight) with increasing pCO2 (Table 2). No significant effects of pCO2 on chlorophyll content were detected for all three species at the end of the experiment (Table 2). Starch content in C. serrulata rhizomes decreased as pCO2 levels increased from 442 to 1204 µatm (Table 2). There were no significant changes for starch content in H. uninervis and T. hemprichii rhizomes with pCO2 enrichment (Table 2). Neither NSC nor soluble carbohydrate content showed significant changes with pCO2 enrichment for all three species (Table 2).

Discussion

Under predicted future scenarios of ocean acidification, marine macrophytes on coral reefs could be amongst the “winners”, because growth and survival will be enhanced by higher CO2 availability (Fabricius et al. 2011; Koch et al. 2013). The present study supports this hypothesis as all three species benefitted from higher rates of photosynthesis (i.e. P max increased) and adjusted photosynthetic kinetics in response to pCO2 enrichment of coastal seawater. Enhanced photosynthetic responses and growth rates were observed after two weeks of exposure to enriched pCO2. Although photosynthetic responses were very similar between species, magnitude of plant-scale responses was species specific.

Physiological responses to pCO2 enrichment

Carbon dioxide enrichment increased net productivity (NP) and energetic surplus (P G:R) in all three species tested. The increase in NP and P G:R was quantified as 0.757–1.040 and 0.322–0.474 units per 100 μatm pCO2, respectively. This response is consistent with previous findings that photosynthetic rates increase with pCO2 in seagrasses (Thom 1996; Zimmerman et al. 1997; Invers et al. 2002; Alexandre et al. 2012). Having greater energetic surplus could indicate flow-on effects to plant-scale responses, such as growth and shoot production (Invers et al. 2002; Palacios and Zimmerman 2007). Energetic status can affect growth, response to physical disturbances such as grazing (Eklöf et al. 2009), abundance and spatial distribution (Dennison et al. 1993; Zimmerman et al. 1997), and even reproductive output (Palacios and Zimmerman 2007).

Maximum photosynthetic rates (P max) and efficiency (α) in all species were raised at higher pCO2 levels although chlorophyll content was not affected. Photoacclimation had been reported for several temperate and tropical seagrass species (Invers et al. 1997; Zimmerman et al. 1997; Jiang et al. 2010; Alexandre et al. 2012), but this is the first study to compare short-term responses to CO2 enrichment amongst three tropical species in one experiment. In general, P–E curves did not show differences between species in their photosynthetic response to CO2 enrichment. The responses over the range of pCO2 in photosynthetic parameters were similar between species (similar slope, as shown by overlapping confidence intervals). Maximum relative electron transport rate (rETRmax) could increase with CO2 enrichment (Jiang et al. 2010). While comparisons between quantum efficiency and O2 production need to be viewed with caution (Beer et al. 2001), these findings indicate that there was a stronger response of P max in the present study using respirometry per 100 µatm rise in pCO2 (7.86 % increase at 1204 µatm for 2 weeks) compared with rETRmax (3.37 % increase in rETRmax at ~807 µatm (pH 7.75) for 3 weeks, Jiang et al. 2010). Temperate species increased P max with pCO2 enrichment too, but showed much more variable response rates per 100 µatm pCO2 (Zostera marina 0.59 % increase per 100 µatm pCO2 for 3 weeks, Zimmerman et al. 1997; Zostera noltii 9.64 % increase per 100 µatm pCO2 for 5 months, Alexandre et al. 2012).

Overall, our study concurs that seagrasses can raise productivity from pCO2 enrichment, at least in the short term (2 weeks exposure). The response of net productivity and P G:R to increasing pCO2 followed a linear trend, indicating that any future change in pCO2 could have an effect on seagrass productivity. In numerous studies on terrestrial plants, the initial stimulation of photosynthesis and growth in elevated CO2 can decline over time, as Rubisco is down-regulated, carbohydrates accumulate and nitrogen content decreases (Stitt and Krapp 1999). Observations of high seagrass abundance at CO2 seep sites indicate seagrass productivity might continually benefit from pCO2 enrichment over the long term (decades) (Fabricius et al. 2011). However, interaction from other co-occurring influences, such as the lowered competition from photosynthetic calcifiers or intrinsic genetic capacity to respond within the population, should be taken into account too. Whether such longer term acclimatory responses to pCO2 enrichment would manifest in tropical seagrasses remains unknown. Furthermore, the capacity of seagrasses to respond to increasing pCO2 is likely to depend on other limiting factors such as nutrient or light availability (Invers et al. 1997).

Light availability is often the primary limiting factor for seagrass productivity. Exposure to low light conditions (such as high turbidity and high epiphyte loads) is a common factor causing seagrass loss (Waycott et al. 2009; Collier et al. 2012). Here, a lowering in the light requirement to meet respiratory demands (E c) and an increase in light efficiency (α) were observed with increasing pCO2. This could imply that a lower amount of light energy would be required to meet metabolic balances (Schwarz et al. 2000; Long et al. 2004). Therefore, CO2 enrichment could potentially increase the tolerance of seagrasses to conditions of low light, for example during flood plume events. In contrast, there was no change in the light level required to reach maximum photosynthetic rates (E k). pCO2 enrichment increased E k in Z. marina (Zimmerman et al. 1997), Z. noltii (Alexandre et al. 2012) and T. hemprichii (Jiang et al. 2010). An increase in P max without a simultaneous rise in light requirement might be explained by the strong upregulation of photosynthetic efficiency (α). pCO2 enrichment can affect light requirements, and this could be important for how seagrasses will respond to changing environmental conditions—including water quality—in the future.

Seagrasses can utilise the predominant \({\text{HCO}}_{ 3}^{ - }\) in seawater via carbon-concentrating mechanisms (CCMs), somewhat alleviating the problem of carbon limitation at higher pH (Durako 1993; Bjork et al. 1997; Uku et al. 2005; Campbell and Fourqurean 2013b). In favourable conditions where other factors are non-limiting, CCMs might cause some seagrasses to be carbon-saturated (Schwarz et al. 2000; Beer et al. 2002). Such mechanisms were thought to be less efficient in Thalassia (T. hemprichii and T. testudinum), rendering this genus less capable of utilising \({\text{HCO}}_{ 3}^{ - }\) than other species (Uku et al. 2005; Campbell and Fourqurean 2013b). Hence, an increase in CO2 availability would be important in raising productivity for Thalassia. Cymodocea and Halodule reportedly possess CCMs that allow them to utilise \({\text{HCO}}_{ 3}^{ - }\) under ambient conditions (Schwarz et al. 2000; Uku et al. 2005). Both species were able to increase photosynthetic rates under enriched pCO2 conditions, where the relative increase in CO2 was much greater than that in \({\text{HCO}}_{ 3}^{ - }\) (Koch et al. 2013). Both species have been observed to become more dominant and have increased biomass around highly enriched volcanic CO2 seeps (Takahashi et al. under review). All the three species tested responded at similar rates in terms of net productivity. It appears that regardless of whether the species possess CCMs or not, CO2 enrichment can increase photosynthetic rates for different species to a similar extent.

Sinks for carbon: plant-scale responses

As a result of increased photosynthetic rates and relatively stable dark respiration rates, energetic surplus (P G:R) was increased at higher pCO2 for all species. The rate of increase with pCO2 levels in P G:R was similar between the three species. There are a number of possible sinks for this additional fixed C. In this short-term study, we measured growth and storage carbohydrates in rhizomes, but other sinks, such as biomass or sexual reproduction, exist.

Response in leaf growth rates to pCO2 enrichment differed between species. Growth of H. uninervis and T. hemprichii responded strongly, but not in C. serrulata. Specifically, relative growth rate (RGR) increased significantly, and in T. hemprichii, an increase in leaf area relative to leaf biomass (SLA) was also observed. Leaf growth response appeared to vary amongst the limited number of studies on tropical seagrass. While Campbell and Fourqurean (2013a) found no differences in leaf growth rates with pCO2 enrichment in T. testudinum, Jiang et al. (2010) showed a 2.63 % rate increase in leaf growth (per 100 µatm pCO2) at pH 7.76 after 3 weeks of exposure in T. hemprichii. This is about half of the 5.62 % rate of increase in leaf growth observed in T. hemprichii here. The effect of CO2 enrichment on growth rate can be influenced by the tissue nutrient requirement of the species and other prevailing environmental conditions (Zimmerman et al. 1997; Palacios and Zimmerman 2007; Jiang et al. 2010; Alexandre et al. 2012; Campbell and Fourqurean 2013a). Under nutrient limitation, seagrasses could direct the fixed carbon towards carbon-rich tissues such as belowground tissues, instead of investing in nitrogen-rich tissue such as leaves (Poorter et al. 1996; Stitt and Krapp 1999). C. serrulata, which has a higher proportion of its biomass existing as shoots and leaves (Hemminga and Duarte 2000), might have required a simultaneous increase in nitrogen availability in order to assimilate the carbon into its leaves. Temperature strongly influences carbon and nitrogen metabolism (Touchette and Burkholder 2007) and could also affect the growth response of seagrasses to pCO2 (Atkin et al. 2005; Collier et al. 2011). Whether these, and other, environmental parameters affected the differences in growth response amongst species warrants further investigation.

Sink strength, or carbon demand, could modulate growth response in seagrasses to CO2 enrichment, similar to that in terrestrial C3 species (Arp 1991; Poorter et al. 1996). Increased energetic surplus from pCO2 enrichment indicates extra assimilated carbon available for storage, growth and metabolism. While little change in NSC content was observed in the present study, seagrasses do possess a number of alternative “carbon sinks”, with the size of carbon demand for each sink dependent on species-specific growth strategy (Doust 1981; Hemminga and Duarte 2000) (described further below). For example, the shorter time taken for shoot initiation for H. uninervis (average 7.9 days), compared to C. serrulata (average 21.2 days) and T. hemprichii (average 38.5 days), meant a faster turnover of aboveground biomass for H. uninervis (Duarte 1991; Marba and Duarte 1998). Therefore, H. uninervis might have a strong carbon demand in leaf growth. The relatively greater proportion of belowground biomass in H. uninervis and T. hemprichii suggests higher storage potential and metabolic demand (Duarte 1991; Marba and Duarte 1998). In these species, more carbon could be directed to belowground biomass, and/or leaf area could be expanded to increase photosynthetic rates. Extra carbon could also be directed to increased shoot production and flowering, as observed in Z. marina after 1 year of CO2 enrichment (Palacios and Zimmerman 2007). Essentially, the extra carbon assimilated could be directed to a single “sink”, such as the growth of new leaves, or it could be spread amongst various metabolic functions and storage organs. The latter makes distinguishing the fate of the extra carbon complicated, especially for short-term experiments such as this study.

In general, our results imply that the availability of higher pCO2 might alter future interspecific competition amongst co-occurring species. With deteriorating water quality, i.e. low light and high nutrients, species that are able to readily assimilate and mobilise carbon resources with the extra CO2 might outcompete other species. Under optimal growth conditions, species that are able to rapidly utilise the extra CO2 to occupy more “space”, i.e. either upwards on vertical stems or via horizontal rhizomes, could potentially increase their abundance and distribution.

Seagrasses as “winners”?

The ability of marine macro-autotrophs to utilise the greater CO2 availability suggests that they will thrive under future scenarios of climate change (Koch et al. 2013). This present study has built evidence to support this, with increased growth, productivity and biomass from pCO2 enrichment (Zimmerman et al. 1997; Invers et al. 2002; Palacios and Zimmerman 2007; Jiang et al. 2010; Campbell and Fourqurean 2013a). This study has also quantified the change in physiological parameters with respect to CO2 enrichment. Surveys at natural CO2 seeps further attest to this, where greater seagrass cover, shoot density, root biomass and productivity were reported at low pH/high CO2 sites when compared to adjacent high pH/low CO2 sites (Hall-Spencer et al. 2008; Fabricius et al. 2011; Russell et al. 2013; Takahashi et al. under review). For calcifying marine autotrophs, such as hard corals, foraminifera and coralline algae, ocean acidification lowers calcification and growth rates and increases rates of bio-erosion (Kuffner et al. 2007; de Putron et al. 2010; Fabricius et al. 2011; Doo et al. 2014; James et al. 2014), and calcifying organisms might be outcompeted (Russell et al. 2011; Short et al. 2014). A shift in the ecological diversity and functions in coastal habitats might result.

This study demonstrated that tropical seagrasses can increase their photosynthetic rates, adjust photosynthetic performance and increase growth rates in response to CO2 enrichment. Varying plant-scale responses to CO2 enrichment between species might affect interspecies competition, especially in mixed species meadows (Takahashi et al. under review). Under CO2 enrichment scenarios, carbon utilisation and allocation traits between seagrass species come into consideration, such as carbon uptake mechanisms, the ability to assimilate additional carbon and the response time of rhizome and shoot elongation to DIC enrichment (Hall-Spencer et al. 2008; Russell et al. 2013; Takahashi et al. under review). Furthermore, environmental conditions such as light and nutrients, which result from water quality changes, could limit species response to CO2 enrichment in the long term. Changes in species composition and diversity in tropical seagrass meadows could potentially impact the functional diversity offered by these productive ecosystems. Interspecific variation amongst seagrasses in response to ocean acidification, over different temporal scales, deserves further examination.