Main

The pH of the oceans will change in the coming decades as the result of anthropogenic emission of carbon dioxide (CO2); changes to ocean chemistry are incontrovertible (Doney et al., 2009). The chemistry is straightforward but not simple; as anthropogenic CO2 increases in the atmosphere, it dissolves in the surface ocean and aqueous carbon dioxide (CO2 (aq)) reacts with water to form carbonic acid (H2CO3)—a weak acid:

Carbonic acid dissociates to hydrogen ions (H+) and bicarbonate ions (HCO3−):

A large fraction of the additional hydrogen ions combines with carbonate ions (CO32−) to form bicarbonate ions:

Therefore, the consequence of dissolving CO2 in seawater is an increase in the concentrations of aqueous carbon dioxide, carbonic acid, hydrogen ions and bicarbonate ions, and a decrease in carbonate ions. As pH is defined by the negative logarithm of the activity of hydrogen ions, it is clear that an increase in H+ must result in a decrease in pH. It is inevitable that the pH of the surface oceans will change with increasing anthropogenic CO2.

Caldeira and Wickett (2003), in their seminal paper, modelled the possible future changes in pH in the oceans and suggested that surface seawater pH would reduce by 0.77 units by the year 2300 if we burned the bulk of available fossil fuels (mostly coal). They also pointed out that, if CO2 emissions continued at the present rate, ocean pH would show larger changes, and faster rates of change, than are detected in the geological record with the possible exception of a few past catastrophic events (Kump et al., 2009). Pearson and Palmer (2000) suggested that the pH of seawater has varied by less than 0.1 units in the past 25 million years. As a result of these and other studies, there has been increasing concern that ‘ocean acidification’ could have severe consequences for marine biota (http://www.ocean-acidification.net). Calcifying organisms, such as corals, molluscs and coccolithophores, will undoubtedly be particularly vulnerable to higher CO2 conditions, as under lower pH the saturation levels decline for calcium carbonate minerals, leading to increased difficulty in maintaining calcite and aragonite shells and skeletons. However, there is little firm evidence for how many other organisms will respond.

All microbes have complex proton pumps that are involved in bioenergetics, but it is not clear how microbes might respond to changes in environmental proton balance. The situation is complicated by the fact that the partial pressure of dissolved CO2 (pCO2) will increase in the future ocean, as well as pH decreasing. Some laboratory studies have shown that higher pCO2 has a fertilizer effect for some phytoplankton species under certain conditions. There is even less information about how heterotrophic microbes might respond to the future, coupled pCO2-pH change in the surface ocean—but there is a great deal of relevant information from other environments that can be evaluated. As ocean warming and increased stratification of the upper water column will also lead to decreases in dissolved oxygen (O2) concentrations (Keeling et al., 2010), changes in the O2/CO2 ratios may impose thermodynamic limitations to aerobic microbial life, especially in regions of low ambient O2 (Brewer and Peltzer, 2009).

Surface ocean pH is variable

Discussions of ocean pH are often made in the context of geological time scales (Pearson and Palmer, 2000; Caldeira and Wickett, 2003), and they imply a constancy and spatial uniformity of pH for the present-day oceans that is inconsistent with field observations. Over timescales of decades or longer, it may be appropriate to consider average annual values. But in reality, pH of the oceans is not constant and there are considerable seasonal, depth and regional variations. Therefore, when considering the overall consequences of ocean acidification for biogeochemical processes, it is necessary to understand that microbes in the present-day ocean experience variable pH; indeed, most of that variability is a consequence of microbial activity. For example, phytoplankton blooms can rapidly reduce pCO2, with a concomitant increase in pH. Although this is the opposite sign in pH change to that expected in the future ocean, it demonstrates that pH is naturally variable and that marine organisms—particularly microbes—must already be capable of adapting to rapid and sometimes large changes in pH. Nevertheless, it is important to know if the average surface water decline of ∼0.3 pH units that is expected by the current century end will present particular challenges for marine microbes or the elemental cycles that they sustain.

Effects of biological activity on ocean pH were recognized in some of the earliest studies of marine chemistry. Atkins (1922) reported that the pH of seawater in laboratory reservoirs was highly variable and that in tanks containing a decaying shrimp, pH decreased from 8.27 to 7.2 over a period of 16 days. Although there may be doubts about the accuracy and precision of these early measurements, it is clear that these changes were a likely consequence of microbial respiration increasing CO2. Rapid changes in CO2, and hence pH, also occur in the present-day ocean. For example, a detailed study in Antarctic coastal water involving daily measurements of pH, clearly demonstrated how pH in the sea is altered by biological activity (Shibca et al., 1977). For most of a 10-month period, pH was 7.9, but in situ acid-base chemistry changed rapidly during a phytoplankton bloom to pH 8.8. Shibca et al. (1977) calculated that this represented pCO2 saturation of only 15%, at a time when the measured O2 saturation was 120%—clear evidence that phytoplankton photosynthesis results in large and rapid changes in CO2 and hence in pH. The potential speed of pH change in a phytoplankton bloom is illustrated in Figure 1. In a mesocosm experiment to investigate ocean acidification (Gilbert et al., 2008), nutrient additions initiated a phytoplankton bloom in an enclosed seawater sample (11 000 l) that had been adjusted to pH 7.8 by bubbling with CO2-enriched air (750 μatm). Within 4 days, through the utilization of CO2, the actively growing phytoplankton assemblage had returned pH to present-day conditions—a change of 0.3 pH units.

Figure 1
figure 1

Rapid change in pH (○) as phytoplankton biomass (▪) increased during a bloom in a mesocosm experiment (Gilbert et al., 2008). Water had been bubbled with CO2-enriched air (750 μatm) and the bloom was initiated by the addition of nitrate and phosphate. The mesocosm was covered and there was no exchange of CO2 with the atmosphere; the observed changes in pH were the result to the utilization of dissolved CO2 by the actively growing phytoplankton assemblage.

Full size image

There are other examples of the importance of biology in changing the surface water pH. Figure 2 compares CO2 at station ALOHA, off the Hawaiian Islands (Dore et al., 2009), with the atmospheric record from Mauna Loa on the Island of Hawaii. Seasonal variations in the surface water are greater than those in the atmosphere because of biological and physical processes (mostly temperature). The pH can change by up to 0.06 pH units during the year even in the oligotrophic Central Pacific, which does not experience the dramatic phytoplankton blooms of temperate oceans.

Figure 2
figure 2

Long-term trends in surface ocean pH and CO2 at station ALOHA in the Central Pacific, along with atmospheric CO2 from nearby Mauna Loa. Net CO2 flux was into the ocean, adding carbonic acid and lowering pH. The long-term declining pH trend was highly significant, but is overlain by substantial seasonal variability (Dore et al., 2009). Direct pH measurements agreed well with calculated values based on measurements of total dissolved inorganic carbon (DIC) and alkalinity Surface ocean pH declined by 0.032 pH units from 1989–2006, an approximate 7% increase in H+ concentration.

Full size image

In many environments, bacteria are already experiencing pH as low (or lower) as that projected for the end of the century in the surface ocean. For example, the thermocline at station ALOHA is a region where respiration exceeds photosynthesis and where sinking organic matter decomposition by aerobic respiration results in the release of CO2 and a reduction in pH (Figure 3). Water below 350 m has a pH less than 7.8, which is projected to be the pH of the surface ocean by the year 2100, yet microbial processes continue. Even in regions of the surface ocean, high respiration:photosynthesis ratios (for example, high latitude regions in winter; residual signal in newly upwelled waters) could result in periodic reduced pH, if dissolved CO2 accumulates. Increases in pH with depth were also noted by Atkins (1922) as a feature of one of the earliest studies of pH in the ocean. Some regions, particularly estuaries and coastal regions, show large seasonal and spatial variations in pH (Salisbury et al., 2008); indeed, they have been among the first regions to demonstrate significant ecological change as a result of a steep decline in seawater pH over time (for example, study of benthic algal and invertebrate community by Wootton et al., 2008).

Figure 3
figure 3

Depth profile of pH at station ALOHA in the central Pacific. The shaded area indicates water that is more acidic than that projected for surface waters at the end of the twenty first century.

Full size image

Given their vital function in marine productivity and planetary habitability, it is imperative to know if microbial assemblages will continue to function at the lower pH values that are projected for the near future. That is, do present-day populations have the metabolic and genetic plasticity to compensate for lower pH conditions? Or are marine microbes less able to acclimate because the average pH of the oceans has changed more slowly over geological time periods, a factor of 30–100 times slower than projected rates for this century, and perhaps has not varied by more than 0.6 pH units for 350 million years (Kump et al., 2009)?

One clue might come from a comparison with freshwater lakes. These have much lower buffering capacity than the oceans, so significant daily variations in pH occur as a result of normal temporal phasing of net photosynthesis and net respiration. For example, Maberly (1996) showed that diel variations in a lake can be as much as 2–3 pH units. In contrast, the pH change in the ocean is expected to be ∼0.3 pH units over the next 100 years. Variations in pH also occur over very small distances. Talling (2006) showed that in some English lakes, pH could change by >2.5 pH units over a depth of only 14 m in the water column. Yet phytoplankton, bacteria, archaea and metazoans are all present in lakes, and appear to be able to accommodate large daily and seasonal changes in pH. Will marine microbes be different, with a lower capacity to acclimate and adapt than freshwater microbes? Is it possible that long-term exposure to much smaller variations in pH means that marine microbes are less able to accommodate pH change than freshwater microbes?

This seems unlikely from the evidence available to date, at least from distribution studies of marine heterotrophic bacteria. For example, a genomic study at station ALOHA (DeLong et al., 2006) investigated the vertical distribution of bacteria. Although there were large differences with depth, particularly, in cyanobacteria and numbers of sequences retrieved, a number of taxa were found throughout the water column to a depth of 4000 m. These included representatives of the alphaproteobacteria and the gammaproteobacteria with 16S sequences from SAR202, SAR324 and SAR406 detected at every depth sampled, including within the core of the acidified zone (200–4000 m). Although it is likely that there were different phenotypes and genetic sub-populations with depth, the 16S sequences were identical (as demonstrated for SAR11 by Field et al., 1997). These data demonstrate that diverse bacterial assemblages, composed of similar lineages, are present over a range of pH (as well as different temperatures and pressures). Important questions that could be readily addressed by sampling a depth profile, or comparing marine, coastal and freshwater systems is: Do marine phytoplankton, bacteria and archaea have the genetic flexibility for homeostasis of intracellular pH (Booth, 1985) under higher CO2/lower pH (that is, they can acclimate)? Or will genetic change be required, involving adaptation by the acquisition of genes through lateral gene transfer or selection of beneficial mutations? Or will small variations in external pH affect chemical equilibria and kinetics at the cell surface and with membrane transporters, leading to consequences for microbial physiology? We urgently need answers to these questions.

Microbially mediated processes—will they be sensitive to ocean acidification?

Table 1 is a brief summary of current knowledge of how key biogeochemical processes may respond in a high CO2 ocean. Experiments to date do not provide a clear indication of how pH change might affect marine microbes and results are inconsistent and at times conflicting. For example, a number of studies have investigated the consequences of reduced pH for calcifying phytoplankton (Table 1). Results by a number of laboratories suggest negative effects of higher CO2/lower pH on coccolithophore cultures (Riebesell et al., 2000, 2007) but at least two studies indicate enhanced calcification under elevated CO2 (Langer et al., 2006; Iglesias-Rodriguez et al., 2008). Ridgwell et al. (2009) have suggested that some of these differences may be due to strain variations in the cultures used for these experiments; or there may be experimental design consequences resulting from the procedures used to adjust the pH (Rost et al., 2008). Clearly, further experiments are required to clarify the effect of ocean acidification on this important group of phytoplankton.

Table 1 Current state of knowledge on microbially mediated process that may be susceptible to ocean acidification
Full size table

Neither are there robust, consistent results on the effects of pH on non-calcifying phytoplankton. Increasing photosynthesis with elevated CO2 is observed for some cyanobacteria (Synechococcus) but not others (Prochlorococcus) (Fu et al., 2007), and many eukaryotic phytoplankton species, most notably diatoms, have carbon concentrating mechanisms that diminish almost entirely the sensitivity of photosynthesis to CO2 variations (Tortell et al., 1997). Bottle incubations and mesocosm experiments with natural plankton communities indicate only a weak sensitivity of primary production to CO2 (Tortell et al., 2000), although limited CO2 fertilization is observed in some cases (Tortell et al., 2008; Egge et al., 2009). There are suggestions that the carbon content of phytoplankton cells may increase under high CO2 conditions (Riebesell et al., 2007), but any physiological changes appear to be quite subtle, and there is conflicting evidence from different studies on how plankton carbon/nitrogen stoichiometry varies with CO2. In terms of all of the major biogeochemical processes listed in Table 1, there is still a state of considerable ignorance about how the ocean system will respond to higher CO2/lower pH.

Future priorities

A number of experimental approaches could be taken that will lead to better understanding. Many of these have been reviewed in the report of a recent expert group of microbial oceanographers that met to consider the consequences of ocean acidification for marine microbes (http://cmore.soest.hawaii.edu/oceanacidification/) and will be only briefly summarized here. A combination of approaches varying greatly in scope and scale was recommended. These range from autecological studies on model marine microbes to investigate intracellular pH regulation, to microcosm and mesocosm scale experiments, and possibly open-ocean, mesoscale perturbation experiments. Comparative studies of similar marine, coastal and freshwater microbial communities (such as those dominated by diatoms or dinoflagellates) could help to answer questions about whether there are intrinsic differences in the response to rapid (that is, hourly or daily) pH change, as occurs in lakes (and to some degree in coastal waters) but not the same extent in open-ocean waters. Much of the present literature on microbe responses to pH/CO2 perturbations is phenomenological, and more detailed studies are required to assess mechanisms at biochemical and cellular levels. Advances in high-throughput sequencing technology now offers opportunities to define how complex microbial communities might respond to CO2/lower pH—for example using 16S rRNA-tag pyrosequencing of the V6 region (Sogin et al., 2006) or similar approaches. Metagenomics and metatranscriptomics technology has now developed sufficiently to study how whole communities might respond to CO2 perturbation (Gilbert et al., 2008). Some existing projects to sample long ocean transects, such as the Atlantic Meridional Transect (Robinson et al. (2006), offer opportunities to test how the increasingly well-defined microbial assemblages in different ocean provinces (Schattenhofer et al., 2009) might vary with pH. Large-scale surveys and time-series efforts are also crucial for scaling up physiological and genomic studies to ecosystems, with the ultimate goal of understanding impacts on biogeochemical cycling and ecosystem services (Doney et al., 2004).

Laboratory experiments with cultured organisms are invaluable for exploring plankton physiological responses to perturbations in pH and CO2. However, it should be recognized and appreciated that most plankton cultures have been maintained in the laboratory for many generations and may not be appropriate model organisms for these investigations on the effects of pH change. Most growth media do not adequately control pH and, at the high cell densities that are common in the maintenance of stock cultures, cells may well have been growing at >pH 9 for phytoplankton, and <pH 6 for heterotrophic bacteria, for many decades in some cases. Existing culture collections may have unwittingly selected for high or low pH conditions, respectively, or for organisms less sensitive to rapid pH variations with time. It may be necessary to conduct experiments using fresh isolates, rather than relying on the more convenient established cultures. Change and selection in laboratory cultures is poorly documented but could be readily investigated. Whole genome sequences are available for an increasing number of phytoplankton and bacterial species. It would be very interesting to compare the genomes of fresh isolates with long established cultures. For example, the common laboratory diatom, Phaeodactylum, was isolated from the English Channel in 1910 (Allen and Nelson, 1910) and has been maintained in culture collections for 100 years—although under poorly defined conditions. How would the genome of this long-term culture compare with a fresh isolate from the English Channel?

Long-term experiments lasting many generations may be necessary to establish how individual organisms might respond to higher CO2/lower pH. Yet these are very challenging experiments. Collins and Bell (2004) maintained Chlamydomonas for 1000 generations under high CO2 (1050 μatm) but failed to find evidence for adaptive, evolutionary change. Long-term adaptation experiments may need to be maintained for decades—much longer than is possible with current funding models (Boyd et al., 2008). And there is no guarantee that adaptation/evolution would be detected at the end of the experiment, so there is a large element of risk in the design and conduct of these studies.

Finally, ocean acidification is only one aspect of habitat variability and climate change. The temperature of the oceans will also increase as a consequence of higher CO2 concentrations in the atmosphere, but little is known about how natural assemblages will respond to higher temperatures. Higher temperatures will also lead to increased stratification that will alter the flux of nutrients from below and the mean light levels experienced by microbes throughout the euphotic zone (Steinacher et al., 2010). Given that ocean acidification will occur at the same time as temperature increases, nutrient decreases and light flux alterations, to name a few habitat variables, there is a strong case for multifactorial experiments to examine the possible synergistic or antagonistic effects of multiple stressors on microbial assemblage diversity and function.

In conclusion, CO2 and pH in the surface ocean are not, and never have been, constant. Microbes in some parts of the present-day ocean already experience average surface ocean pH that will occur by the end of the century. And freshwater and coastal microbes experience short-term and seasonal changes that are many times greater than those that will occur in the open-ocean. Given these facts, perhaps the most appropriate null hypothesis to test is that marine microbes possess the flexibility to accommodate pH change and there will be no catastrophic changes in marine biogeochemical processes that are driven by phytoplankton, bacteria and archaea. Clearly, calcifying organisms are a special case as carbonate minerals will be less saturated—and for the case of aragonite, undersaturated in surface waters in a high-CO2 ocean. Photosynthetic organisms may also be influenced and it is even possible that higher CO2 may be beneficial. But the rest of the microbial community should not be assumed to be at risk until evidence to the contrary is obtained.