Over the last century, humans have produced large amounts of CO₂ through activities such as fossil fuel burning, intensive agriculture and deforestation [1]. The concentration of CO₂ in the atmosphere has risen from 280 ppm in preindustrial times to 400 ppm in 2015, as measured at Mauna Loa Observatory, which holds one of the longest records for direct measurement of atmospheric carbon dioxide [2]. By the middle of this century, atmospheric CO₂ levels could reach more than 500 ppm and surpass 800 ppm by the end of the century [3]. During the industrial era, the ocean has absorbed about one-quarter of this anthropogenic CO₂ [4,5], acting as a carbon sink and thus contributing to the key ecosystem service of climate regulation. A portion of the CO₂ absorbed by oceans is stored in living biomass and sequestered in sediments but a large amount remains in its inorganic form. This increase of inorganic carbon stored in the oceans has driven a reduction in seawater pH and promoted changes in the seawater chemistry in a process commonly referred to as “ocean acidification” (OA) [6–10]. The global ocean pH is expected to fall to between 8.05 and 7.6 by the end of this century (IPCC, 2013; [11]). This raises concern about the possible impacts of these changes on marine organisms. OA is a ubiquitous stressor, which is predicted to lead to negative consequences in the future for marine organisms [12], ecosystems [13] and the provision of ecosystem services [14]. In recent years, studies have underscored the crucial role of shallow coastal ecosystems, which function not only as transition zones between land and ocean but also as filters for carbon sequestration [15,16]

Seagrasses are marine flowering plants that form one of the richest coastal ecosystems [17], providing a large number of ecological services [16–18] including long-term carbon catchment [16,18]. They cover less than 0.2% of the ocean surface but make a disproportionately high contribution to marine net primary productivity (NPP) (1%) and are responsible for approximately 15% of the carbon stored in oceans [19]. Seagrass-dominated ecosystems are highly productive habitats [20] that play an important role in the carbon cycle of coastal areas [20,21]. The excess organic matter that they produce can be exported to adjacent ecosystems in particulate or dissolved forms [22,23]. Dissolved organic carbon (DOC), with about 700 PgC, represents only 2% of the carbon pool in the ocean [24]. However, it is a central factor in the global carbon cycle [25,26], acting as a vector for quick transfer of C and energy in food webs, as it is easily assimilated by marine organisms and fully involved in the C exchange between communities [25,27]. The global net DOC exported from seagrass meadows, as calculated by Barrón et al [28], represents as much as the 46% of the global NCP of seagrass meadows, as reported by Duarte et al [29]. Thus, this released DOC may be critical in maintaining the high annual productivity of communities dominated by seagrasses [30]. However, few studies have focused on the relation between productivity and DOC flux, especially in situ (for review, see Barrón et al [28]). While there has been extensive research in the DOC dynamics of the open ocean (for reviews, see Hansell & Carlson [25,31]), the role of coastal ecosystems in the global DOC cycle is still inadequately understood, despite being a strong sink for ocean DOC [32].

Increased anthropogenic pressure has led to widespread loss of seagrasses in shallow estuarine and coastal zones [33]. With the decline in the global surface area of meadows measured at 7% yr^-1 [34], seagrass meadows are the most quickly declining ecosystem on the planet. OA could either ameliorate or aggravate this current decline in seagrasses [10]. Thus, the response of seagrasses to OA must be considered for effective management of coastal regions in the future. The effects of increased CO₂ concentrations on seagrasses may depend on the degree of carbon limitation in natural systems. Carbon limitation has been partly attributed to the thickness of the diffusive boundary layer (DBL) surrounding leaf surfaces at low current velocities [35–37] or to a relatively inefficient HCO₃^- uptake system [38]. Increases in water current velocity are linked to a thinner DBL [39] and consequently, faster transfer of CO₂ molecules from the water column to seagrass cells. As higher current velocity leads to a reduction in the DBL thickness, this favours CO₂ uptake and photosynthesis [40]. However, increasing flow velocity may reduce photosynthetic rates due to enhanced sediment resuspension (i.e. decreasing light levels; [41]) and increased self-shading by leaves, because leaves tend to collapse onto each other when exposed to strong currents [40]. Hence, current velocity conditions can result in a net positive or negative effect on seagrass productivity, depending on which effect is stronger: an increase in CO₂ uptake as a result of the decrease in DBL or a reduction in photosynthesis as a consequence of sediment resuspension or leaf self-shading. It is important to highlight here that hydrodynamic conditions in coastal areas worldwide may also change as a consequence of anthropogenic engineering activities that change tidal flows [42] and climate change, which is expected to increase wave stress, and the frequency and intensity of storms [43].

Studies on the effects of OA on seagrasses so far have mainly focused on how elevated CO₂ concentrations will affect the productivity, light requirements and nutrient content of seagrasses [44–47]. However, the net impact of higher CO₂ is still unclear because there are many ambient variables that could ameliorate or aggravate the effects of CO₂ increase. The interactions between an increase in CO₂ and hydrodynamic conditions have not been addressed in any depth, in spite of the importance of the latter at different levels in seagrass ecosystems [48–51]. Likewise, little attention has been given to the effect of hydrodynamics on DOC flux in seagrass meadows. Recent studies have highlighted the direct relationship between ecosystem productivity and DOC flux [28]. Therefore, if hydrodynamic conditions affect photosynthesis, and thus productivity [40], current velocity may also have a significant effect on the DOC flux released by seagrass populations, which in turn could affect the flux of C in coastal communities. Therefore, our aim in this study is to explore the interactive effects of ocean acidification and water current velocity on productivity and the release of DOC in the temperate seagrass Zostera noltei.

The two factors assayed (pH and current velocity) produced significant changes in the carbon metabolism and dissolved organic carbon (DOC) flux in Z. noltei. Lower pH mainly enhanced seagrass productivity, whereas current velocity mainly enhanced DOC flux. Throughout the experiment, Z. noltei remained net autotrophic, independently of the treatment condition, and maintained high productivity under all the conditions tested. On average, GPP was 65 ± 7 μmol C gFW^-1 d^-1, which is within the range of values described by Duarte et al [29] for a large number of seagrass species. A significant increase (150%) in GPP and R under the more acidic forecasted pH conditions was found. This may indicate that photosynthesis of Z. noltei is carbon limited at the current inorganic carbon concentration of seawater (i.e. CO₂), as found in previous studies on this species (e.g. [65,66]), and also in studies on other seagrass species [44,45,67]. Hence, Z. noltei may benefit from future CO₂ enrichment through the enhancement of photosynthetic rates at higher CO₂ concentrations [65]. These higher CO₂ concentrations may lead to significant higher uptake of C, which could in turn be stored in the tissues of Z. noltei (e.g. increase in non-structural carbohydrates reserves as found in this study) or exported in particulate or dissolved form (as also found in this experiment), contributing to an increase in the productivity of adjacent communities.

We acknowledge that the presented experimental design, although have been widely used in the literature for ocean acidification research (see Cornwall & Hurd 2016 [68]), can be considered inappropriate according to Humbert 1984 [69]. The six treatments used here were independent respect the factor current velocity (i.e. each unit had their own flume tank and water pump) but inter-dependent respect the factor pH, as all units of each pH conditions were feed with the same large reservoir (i.e. one header tank of seawater for CpH and another for FpH where the CO₂ was manipulated). Then, a single chance of motor failure (e.g. water contamination event or other kind of intrusion) could have produced a “spurious treatment effect”. Experimental logistics prevented the use of a pH header tank for each unit. However, the use of two large reservoirs that feed all replicates of each pH treatment aided to reduce experimental variability and to maintain the required water conditions for a longer time (i.e. water temperature, alkalinity, etc.), as demonstrated by our daily monitoring of water physical-chemical variables (see results section, abiotic variable subsection and Table 2). In addition, each unit of each pH treatment was manipulated as an independent system (i.e. with their own independent maintenance system, their own water pump etc.) and all experimental units in one pH treatment was equally exposed to the same conditions as all experimental units in the another pH treatment, which reduced the possible bias according to Hurlbert 2013 [70]. On the other hand, the rate of water input into the two large reservoirs and in each units were very high (~300 l d^-1 for large reservoirs and ~50 l d^-1 for units). Hence, microorganisms in water do not biologically modify seawater carbonate chemistry significantly over the duration of the experiment as it noted in Table 1. Therefore, the possibility of “spurious treatment effect” as a consequence of experimental design, if occurred, should have been very low and then, the variability associated with this experimental design does not mask the main results of the research.

Current velocity did not have significant effects on plant productivity, although plants subject to MV under CpH conditions had higher GPP (170%) than the average of the other two velocity levels. Under low current velocities, the thickness of the DBL surrounding the leaf surface likely increases, which reduces carbon uptake, as previously demonstrated for inorganic carbon [35–38] and other nutrients [48,71,72]. In contrast, under high current velocities, self-shading in the population increases since leaves orient themselves almost horizontally in the strong current, collapsing onto each other and affecting light absorption and thus productivity [40]. The intermediate current velocity likely allows for a more favourable balance between DBL reduction and light absorption, leading to an increase in plant productivity. Some previous studies conducted in the same small flumes have shown the opposite trends (mainly at high current velocity) to those found in this experiment [51,53,72]. For example, in these previous studies, higher current velocities actually favoured the growth and development of Z. noltei plants, mainly because of the more horizontal position of their leaves, which enhanced light capture. The relatively lower plant growth under MV was explained by insufficient oxygenation of the rhizosphere, which requires energy to offset the effects of the anoxic sediment [73–75]. However, the main difference between our experimental design and that of these previous studies is shoot density. We used a high shoot density (c.a. 5,650 shoots m^-2) in our design, whereas shoot density was one order of magnitude lower (c.a. 150 shoots m^-2) in these previous studies. Under such experimental conditions, the interaction between density and current velocity can produce opposite effects. Under higher shoot density and MV conditions, the associated higher complexity of the belowground network in Z. noltei beds may lead to greater oxygenation of the rhizosphere; under HV conditions, light capture may decrease because of increased self-shading by leaves in the canopy. The interactions between pH, current velocity and shoot density deserve further research. However, the differences in GPP between different levels of current velocity disappeared under FpH, which suggests that Z. noltei plants were CO₂ saturated under more acidic conditions. Thus, CO₂ enrichment may limit the impact of light limitation on seagrasses [76] by offsetting the lower light absorption under HV conditions, while also probably offsetting the increase in DBL thickness surrounding the leaf surface under LV conditions. Thus, the interactive effects of pH and current velocity on the productivity of Z. noltei populations are not as straightforward as expected: under current levels of pH, current velocity may have a positive effect depending on shoot density; under the lower pH levels forecast for the future, this positive effect of current velocity may disappear (Fig 5).

Regarding the effects on non-structural carbohydrates, sucrose increased in aboveground tissues under FpH conditions, while a non-significant response was found in belowground sucrose and in both above- and belowground starch content. However, belowground starch increased by 30–260% in plants subjected to the treatment conditions of LV + CpH (Fig 3 and Table 4). This is consistent with previous short-term experiments on the effects of OA in seagrasses, which showed an increase in the concentration of non-structural belowground carbohydrates [47,77]. Carbon reserves are essential for seagrass survival under stressful environmental conditions because environmental stressors, such as periods of light limitation [78–81], ammonium toxicity and/or sediment anoxia [64,82], increase plant demand for carbon. Under these stressful conditions, plants are forced to use stored carbohydrates to maintain their carbon skeletons and provide energy during these periods of low C fixation through photosynthesis.

One of the most noteworthy results of this study was the positive relation between DOC flux and current velocity, independently of pH level. It is known that the DOC released to the ocean depends on many factors (e.g. temperature, light, community respiration, nutrients, etc. [25, 28]). The present study demonstrates that among these factors, hydrodynamic conditions is an important factor determining the dynamic of DOC in seagrass communities (Figs 4 and 6 and Table 3). Our results show that circa of 50% of the variability in DOC release in our experimental seagrass community was explained by hydrodynamic conditions (Fig 6). The remainder may be explained by the existence of other factors that are also known to affect the release of DOC in seagrass communities such as temperature, light, nutrients, bacterial use of DOC, etc. The positive relation between DOC flux and current velocity can be attributed to the sum of two different processes. First, as demonstrated by this work, increasing current velocity results in higher GPP of Z. noltei population, and since a part of the GPP is released as DOC [28], increasing current velocity indirectly results in a higher release of DOC. Second, seagrass ecosystems store large amounts of particulate organic carbon in the sediment [83], which can potentially be transformed into DOC through the activity of heterotrophic bacteria [84,85]. Although the relative contribution of each compartment (e.g. DOC released from sediment) to the net DOC flux in seagrass communities remains poorly understood [28], higher current velocity is expected to also enhance the transfer of DOC from the sediment to the water column, as demonstrated in other compounds [40,86–88]. This finding has interesting ecological implications since the DOC released from seagrass populations means quicker and more efficient transfer of carbon and energy from primary producers to higher trophic levels; thus, seagrass populations thriving in areas with higher hydrodynamics may produce higher DOC fluxes.

Besides seawater acidification and current velocity, other factors linked to climate change, such as warming, may severely impact the productivity and even survival of marine organisms [89]. By the end of this century, mean global surface temperatures are projected to increase by ~3–4°C [90]. Temperature is a key factor for seagrass health, growth and community metabolic rates [10]. Thus, increases in temperature may exacerbate or alleviate the results found here. For example, the Metabolic Theory of Ecology (MTE) predicts that respiration rates will increase more quickly under warming scenarios than primary production rates [91–93], which results in a reduction of the P:R ratio and a greater chance of shifting systems into heterotrophy. On the other hand, recent studies have shown that the net DOC flux in seagrass communities are positively correlated with water temperature [23,28]. Hence, future research should delve into the interactions between ocean acidification and current velocity with other environmental or anthropogenic stressors linked to climate change, such as warming, in order to understand and mitigate the effects of this global threat on seagrass ecosystems.

In summary, the best scenario in this study for Z. noltei was under OA and high current velocity, which led to a significant increase in productivity and non-structural carbohydrate concentrations, especially in aboveground sucrose and starch, and the lowest levels of biomass loss. In addition, a direct relationship between current velocity and DOC flux was recorded at the end of the experiment under both OA and control conditions. Since both factors (i.e. OA and hydrodynamics) are expected to increase under climate change, and without considering other indirect interactions such as warming, sea level change, herbivory, etc., seawater acidification and higher current velocity may produce a favourable scenario for Z. noltei populations, increasing their productivity, their non-structural carbohydrate levels and release of DOC, which means a quicker and more efficient transfer of carbon and energy to higher trophic levels, and higher resistance and resilience to external stressors.