In the 20th century, the atmospheric carbon dioxide (CO2) concentration continued to increase as a result of fossil fuel combustion and other human activities. It was in turn taken up by the ocean gradually through air–sea exchange. Oceanic CO2 can hydrolyze to increase the concentration of hydrogen ions (H+), which leads to the reduction of pH in the ocean by 0.1 units (Orr et al., 2005). Based on the amount of global CO2 emissions at present, the pH of the ocean is likely to drop by 0.3–0.4 units by the end of the 21st century and by 0.7 units after 300 years. The current and predicted levels of CO2 and acidity of seawater of nearshore, estuarine, and higher-latitude habitats are expected to be even greater and substantially more variable than those of the open ocean (Gruber et al., 2012; Zhai et al., 2014). One alarming consequence is the rapid change in seawater chemistry and decrease in ocean pH, which could have great impacts on marine ecosystems and pose a threat to marine life (Frommel et al., 2012; Kerr, 2010). Elevated CO2 concentrations can disturb the acid-base regulation, blood circulation, respiration, and the nervous system of marine organisms, leading to long-term effects such as reduced growth rates and reproduction (Frommel et al., 2012). Other direct response to ocean acidification (OA) have been found in the alteration of behaviour (Dixson et al., 2010; Munday et al., 2009a), development (Frommel et al., 2013), RNA/DNA ratio (Franke and Clemmesen, 2011), and otoliths (Checkley et al., 2009; Maneja et al., 2013; Munday et al., 2011b) of marine fish larvae. However, the emerging picture remains intriguingly complex. While the majority of responses to high CO2 appear to be negative (Branch et al., 2013) with highest sensitivities observed during the early life stages (ELS) and in the calcifying invertebrates such as corals, bivalves, pteropods, and echinoderms, there are substantial evidences for non-linear, neutral, or even positive reaction to increasing CO2 conditions (Hurst et al., 2013; Munday et al., 2011b; Murray et al., 2014). Moreover, marine fish exemplifies this complexity. Decades of empirical data suggest that juvenile and adult fish possess sufficient acid-base and osmoregulatory capabilities for the toleration of very high metabolic and ambient CO2 levels (> 2000 µatm) (Murray et al., 2014). Although fish possess the ability of acid-base balance regulation, its physiological function will certainly decline under such regulation for a long time from the perspective of energetics, especially in its most fragile and sensitive early life stage during its life history. In addition, in ELS of multiple taxa including fish, elevated CO2 was shown to affect calcification of shells and skeletons due to a drop in the carbonate availability (Riebesell, et al., 2010). Munday et al. (2011a) observed no effect on spiny damselfish otolith calcification at 850 µatm, while Munday et al. (2011b) and Checkley et al. (2009) highlighted an otolith hypercalcification in white seabass (Atractoscion nobilis) larvae exposed at 993 and 2558 µatm pCO2 and in clownfish (Pomacentridae) larvae at 1721 µatm pCO2, respectively. In case of calcification modulation, otolith morphology can be affected, which may have negative repercussions on the behaviour and acoustic function of fish and decrease their survival probabilities (Bignami et al., 2013; Réveillac et al., 2015).

Marine medaka, Oryzias melastigma or Oryzias javanicus, is one of the 14 species belonging to the genus Oryzias, which distribute in estuarine waters from East to Southeast Asia (Koyama et al., 2008). It has been proposed as a model species in marine environmental risk assessments (Mu et al., 2014). However, few studies have addressed OA effects on the ELS of marine medaka so far. The objective of this study was to examine how CO2-driven OA affected the embryos and newly hatched larvae of marine medaka after 21 days exposure through investigating the embryonic, larval, and otolith development.

Assessment of species sensitivity or tolerance to CO2-driven acidification in marine environment is critical to evaluate the impacts of OA on marine biodiversity and ecosystem function (Fabry et al., 2008; Melzner et al., 2009). A number of studies found that CO2-driven acidification had obvious influences on ELS of many marine invertebrates, especially calcified organisms including coral (Doropoulos et al., 2012; Fabricius et al., 2011), coccolithophores (Berry et al., 2002), and mollusk (Kroeker et al., 2013; Thomsen et al., 2013; Waldbusser et al., 2011). OA was predicted to potentially affect individual behaviour such as development, growth, survival, and swimming particularly during the early life stage of marine organisms (Munday et al., 2008). In our experiments, the duration of embryonic stage, egg survival and embryonic heart rate of marine medaka were unaffected by acidification water with pH 7.6 and pH 7.2. There was a slight increase in the embryonic duration of the eggs, but the size effect was not different among the three pH treatments. Overall, these results suggest that the egg stage of marine medaka is relatively tolerant to elevated CO2 and low pH level, which were consistent with the results reported by other studies on a diverse set of marine fishes. For instance, Munday et al. (2009b) found the survival to hatch of orange clownfish (Amphiprion percula) from the Great Barrier Reef, Australia, to be nonresponsive to pCO2 levels to 1020 ppm (pH 7.8). Similarly, Franke and Clemmesen (2012) found no significant effect of elevated pCO2 levels from 460 to 4635 ppm (corresponding to pH 8.08–pH 7.05) on survival to hatch of Atlantic herring from the western Baltic Sea. In the study of Frommel et al. (2013), the survival of embryos of Atlantic cod from the Bornholm Basin of the western Baltic Sea was not altered at pCO2 levels up to 4000 ppm (pH 7.2). Hurst et al. (2013) also reported no effect on embryo survival of walleye pollock (Theragra chalcogramma), common in the temperate eastern North Pacific, at pCO2 levels up to 1933 ppm (pH 7.4). In other cases, however, a strong effect of CO2 was evident on the embryo survival of summer flounder (Paralichthys dentatus), an ecologically and economically important flatfish of the inshore and nearshore waters of the mid-Atlantic Bight (Chambers et al., 2013). The relative survival of summer flounder embryos was reduced to 48 % when maintained at 1808 ppm pCO2 (pH 7.5) and to 16 % when maintained at 4714 ppm pCO2 (pH 7.1). Baumann et al. (2012) also reported a 74 % reduction in survival of embryos and young larvae of inland silverside, Menidia beryllina, native to estuaries of the US Atlantic coast, when maintained at 1100 ppm pCO2 compared to those held at 410 ppm pCO2. All of these studies varied in the number of parents used, the time lapse between egg fertilization and initiation of CO2 treatment, and how and when survival was scored. For example, the CO2 treatments of inland silverside by Baumann et al. (2012) began at approximately 24 h post-fertilization, and the survival was scored at approximately 1 week post-hatching. The different approaches used in previous studies may preclude a fair cross-study comparison (Chambers et al., 2014); however, the overall present of effect of elevated CO2 environments on embryo survival is in contrast to the findings here. Habitats occupied of species, particularly in the ELS, may play a role in their sensitivities. It is counter to expectations and requires further attention that species in their ELS are found in estuarine (marine medaka) and inner shelf (summer flounder) habitats, both with relatively high ambient CO2 levels, but exhibit different sensitivities to experimentally elevated-CO2 levels.

An unexpected result of our study was that elevated levels of CO2 affected larval development abnormalities, and the average deformity rate of marine medaka larvae (approximately 1 week post-hatch) increased significantly by 16 % as CO2 increased from control level (pH 8.2) to high-CO2 level (pH 7.2). Although CO2-induced acidification up to the high-CO2 level (pH 7.2) had no noticeable effect on larval survival by the end of the experiments (21 days), the larval development abnormalities may ultimately influence the later life consequences and therefore further reduce the productivity of fish stock in future acidified oceans. Chambers et al. (2014) found no reduction in survival with CO2 for larvae during the first 4 weeks of larval life (experiment ended at 28 days post-hatching (dph)), however, the sizes, shapes, and developmental status of larvae showed initially longer and faster growing when reared at pH 7.5 and pH 7.1 levels, and the tissue damage was evident in larvae as early as 7 dph from both elevated-CO2 levels. At present, it is unknown how increasing CO2 levels affect development and survival in fish ELS. Even if fish embryos and early larvae are capable of physiological adaptation to increased CO2 somehow, this would incur further metabolic costs and thus reduce energy available for tissue synthesis or post-hatch survival on diminished yolk reserves. As some fish eggs, including those of O. melastigma, seem to be tolerant to low-pH conditions, the high levels of CO2 or associated changes in carbonate chemistry may be more important to larval-fish development than hydrogen ion concentrations. (Baumann et al., 2011; Ishimatsu et al., 2008).

The pH drop driven by CO2 can change concentrations of bicarbonate and non-bicarbonate ions during which elevated CO2 affects saturation states of calcium ions carbonate polymorphs (Munday et al., 2008). Otoliths are bony structures of fish to sense orientation and acceleration and consist of aragonite-protein bilayers, which document fish age and growth (Checkley et al., 2009). Its formation starts during embryonic development, and any alteration of otolith size or shape is important for physical performance and individual adaptability of fish. Therefore, any substantial change in the size, shape, or symmetry of otoliths could have serious implications for individual performance and survival (Munday et al., 2008, 2011a). In this study, we found no significant difference existing between the left and right sides of marine medaka larval otolith under the same pH level. However, otolith area of larval fish exposed to the intermediate-CO2 level (pH 7.6) was smaller than that of control. Results suggested that there was no significant pCO2 effect on otolith symmetry of marine medaka, defined as the difference between the right and left sides. However, the otolith area was significantly affected. The trend of reduction in otolith area of marine medaka larvae exposed to elevated CO2 environments found here has not been reported in most previous studies focusing on other marine fishes. For instance, Checkley et al. (2009) found that otolith area of white seabass (Atractoscion nobilis) larvae increased by 7–9 and 10–14 % after exposure to 993 and 2558 ppm CO2, respectively. Munday et al. (2011b) found that the size, shape, and symmetry of otoliths in larval clownfish was unaffected by exposure to simulated levels of OA (pH 7.8 and 1050 µatm CO2); however, in a more extreme treatment (pH 7.6 and 1721 µatm CO2) otolith area and maximum length were larger than those of control otoliths. Maneja et al. (2013) found that elevated CO2 had no significant effect on the shape of the otoliths nor was there any trend in the fluctuating asymmetry, while increased otolith growth was observed in 7 to 46 d post hatch cod larvae in two pCO2 treatments of 1800 and 4200 µatm. In contrast, Munday et al. (2011a) did not detect any effect of elevated CO2 on otolith size of juvenile spiny damselfish, Acanthochromis polyacanthus, which were reared for 3 weeks in treatments up to 841 µatm CO2. Our results seemed to support the hypothesis that otoliths of larvae reared in seawater with elevated CO2 would grow more slowly than they do in seawater with normal CO2. The reduction of otolith area was likely associated with reduced CaCO3 saturation which slowed down its formation. We do not know whether smaller otoliths have a deleterious effect, although we do know that asymmetry between otoliths can be harmful (Checkley et al., 2009).The difference between our results and other studies may be related to: (1) different pCO2 levels; (2) different life histories; or (3) different exposure duration (Munday et al., 2011a). However, another interesting result from the present study was that the otolith area of marine medaka larvae under the extreme CO2 level (pH 7.2) tended to increase instead of continuously reduce. We should not ignore its own acid-base regulation ability that increased the available amount of carbonate by compensation mechanism for otolith to intensify the calcification process under such acidic condition (Checkley et al., 2009). Calcium incorporation into the otolith was modulated by the seawater pH. This questions the stability of the element:Ca ratio under environmental hypercapnia. During the biomineralization of the otolith, chemical elements such as metals and metalloids are supposed to substitute for calcium (Réveillac et al., 2015). The changes of pH and seawater chemistry caused by increased CO2 can modify the speciation of metals and their subsequent bioavailability to organisms (Millero et al., 2006). The physiological response of fish to hypercapnia might in turn stimulate processes to compensate for acidosis based on the key role of ion transporters. In the present study, OA may interfere with trace element uptake and body concentrations and therefore could affect otolith growth and microchemical constituent. Further studies are thus needed to investigate the possibility that OA impacts on the trace metals properties, molecular-binding affinities and incorporation pathway into the otolith.

In conclusion, this study demonstrated that, under projected near-future pCO2 levels, the ELS of marine medaka exhibited a dramatic increase of larval developmental deformity and otolith calcification while their survival was not affected. Importantly, the observed CO2-induced abnormal development of larvae might have predictably negative consequences on the recruitment of fish population, the effects of which on later life history and phenotype of subsequent generations should be concerned. As the otolith is an essential tool used in reconstructing fish life history in terms of age, somatic growth and attended habitats, further studies should investigate the process of otolith biomineralization. Finally, we emphasize that there is considerable variation among species in their sensitivities to elevated CO2 and reduced pH. Determining the traits that render some species more susceptible than others will be helpful and valuable in predicting the long-term and ecological effects of OA.