Resolving the interactions of ocean acidification and temperature on coral calcification media pH

Ocean acidification typically reduces the calcification rates of massive Porites spp. corals, but increasing seawater temperatures (below the stress and bleaching threshold) can offset this effect. Here, we use δ11B to reconstruct the pH of the calcification media (pHECM) used to precipitate the skeleton in poritid corals cultured over a range of seawater pCO2 and at 25 °C and 28 °C. Increasing temperature had no significant effect on pHECM at high pCO2 although corals increased their calcification rates. pHECM was reduced at 28 °C compared to 25 °C at low seawater pCO2, although calcification rates remained constant. Increasing calcification rates could reflect the positive influence of temperature on aragonite precipitation rate, an increase in calcification media saturation state or a change in the concentration/behaviour of the skeletal organic matrix. The two temperatures utilized in this study were within the seasonal range at the coral collection site and do not represent a heat stress scenario. Increasing seawater temperatures may promote calcification in some corals in the future but are unlikely to benefit the majority of corals, which are already living close to their maximum thermal tolerance limits.


Introduction
Coral reefs are amongst the most diverse ecosystems and support the livelihoods of millions of people (Cinner 2014). Rising atmospheric CO 2 is increasing the temperatures and decreasing the pH of most surface seawaters (IPCC 2014). Understanding how these changes affect the accretion of aragonite skeletons by the tropical corals that underpin reef structures is key to predicting the future of coral reefs. Coral calcification is usually reduced at lower seawater pH (Erez et al. 2011), but massive Porites spp. corals, important reef-building species in the Indo-Pacific (Veron 1993), appear to be relatively resilient to ocean acidification (Fabricius et al. 2011;Crook et al. 2012). In culture studies, the calcification and linear extension rates of this genus can be reduced at high seawater pCO 2 (Cole et al. 2016), but this suppression can be mitigated by increases in seawater temperature below the temperature stress threshold (Anthony et al. 2008;Edmunds 2012;Cole et al. 2018). The response of calcification to temperature change is seawater pCO 2 dependent (Anthony et al. 2008;Cole et al. 2018;Edmunds 2012). The precipitation rate of synthetic (inorganic) aragonites is positively correlated with temperature (Burton and Walter 1987). However, the coral calcification increase observed by Cole et al. (2018) at high pCO 2 is much higher than that predicted from synthetic aragonites, while at low pCO 2 , the change in coral Topic Editor Morgan S. Pratchett calcification rate is lower than that predicted from synthetic aragonites. Understanding how temperature and seawater pCO 2 interact to drive these changes in coral calcification is a goal of the current study.
The coral skeleton forms from calcifying media (ECM) which may be contained between the basal layer of the coral tissue, the calicoblastic epithelia, and the underlying coral skeleton  or contained within vesicles in the coral tissue (Mass et al. 2017). This calcification media is semi-isolated from ambient seawater, and both the pH and dissolved inorganic carbon (DIC) of the media are significantly higher than in local seawater (Sevilgen et al. 2019), facilitating precipitation of the aragonite skeleton. The three predominant forms of dissolved inorganic carbon in seawater are CO 2 , bicarbonate (HCO 3 -) and carbonate (CO 3 2-). Increasing seawater pH shifts the DIC equilibrium to reduce the proportion of CO 2 and increase the proportion of CO 3 2-. CO 2 can readily diffuse across cell membranes so the pH increase facilitates the invasion of CO 2 into the calcification site and acts as a DIC concentration mechanism (Erez 1978). This allows the [DIC], [HCO 3 -] and [CO 3 2-] to increase above that of seawater providing the DIC required to build the skeleton (Allison et al. 2014).
Here, we use skeletal d 11 B to estimate the pH of the calcification media (pH ECM ) in a series of corals cultured over a range of seawater pCO 2 (180, 260, 400 and 750 latm) and at 25 and 28°C ). These pCO 2 conditions reflect conditions in the Last Glacial Maximum, the pre-industrial (Gattuso and Lavigne 2009), the present day and a future scenario. B speciation in seawater is controlled by ambient pH, and there is a large B isotope fractionation between the predominate dissolved species with borate, B(OH) 4 -, depleted in 11 B compared to boric acid, B(OH) 3 (Kakihana et al. 1977). Assuming that B(OH) 4 is predominantly incorporated into the lattice during aragonite precipitation (Sen et al. 1994;Noireaux et al. 2015;Balan et al. 2016), skeletal d 11 B reflects pH ECM . Reports of B(OH) 3 in tropical (Klochko et al. 2009) and deep water corals (Rollion-Bard et al. 2011) may reflect a change in B coordination (from tetrahedral B4 to trigonal B3) after aragonite precipitation (Klochko et al. 2009;Noireaux et al. 2015;Balan et al. 2016) or the incorporation of B3 in coral centres of calcification (Rollion-Bard et al. 2011). In this study, we avoid the analysis of centres of calcification and we assume that coral skeletal boron is derived from dissolved B(OH) 4 -. It is unclear how avoiding these areas affects our estimates of skeletal d 11 B. Centres of calcification exhibit different crystal morphologies (Cohen et al. 2001), contain higher concentrations of organic materials (Cuif et al. 2003) and exhibit different geochemistry (Allison 1996) compared to the remainder of the skeleton. However, centres of calcification occupy a small volume of the total skeleton (* 4%, Allison et al. 2005), so the effect of exclusion of this material from analysis is likely to be small. We determine how pH ECM is affected by seawater pCO 2 and temperature in four genotypes of massive Porites spp. and we explore how pH ECM relates to published calcification rates for these corals ).

Methods and materials Coral culturing
We cultured multiple genotypes of massive Porites spp. corals over a range of seawater pCO 2 (* 180, 260, 400 and 750 latm) and temperature (25 and 28°C). Large heads of massive Porites spp. corals were collected in Fiji and imported into the UK. Heads were identified to species level based on corallite morphology (Veron 1993) and were considered to represent different genotypes when they were collected from spatially separate (non-adjoining) colonies. The physiological responses of small experimental corals are not representative of larger colonies (Edmunds and Burgess 2016), and we sawed imported heads into multiple pieces (each * 12 cm in diameter) so that at least one large piece of each genotype could be cultured in each seawater pCO 2 treatment.
Corals were cultured in seawater bubbled with the target air:CO 2 gas mixes in a purpose-built large-volume aquarium system (Cole et al. 2016. For each pCO 2 treatment coral pieces were split randomly between 2 tanks of 21 l each, fed with seawater recirculated from a highdensity polyethylene reservoir containing * 900 l of seawater. Only one reservoir was used for each pCO 2 treatment. To test for variations between replicate reservoirs results from this study were compared to data collected from a previous study using the same aquarium system and culturing corals over a range of seawater pCO 2 at 25°C . After import, corals were maintained at ambient seawater pCO 2 conditions for 2 months, kept under varying CO 2 for 1 month, while the seawater pCO 2 was gradually adjusted to the treatment conditions and then acclimated at the final treatment pCO 2 for 4 months at 28°C. Colonies were then stained with Alizarin Red S and cultured for a 5-week experimental period during which calcification, respiration and net and gross photosynthesis were measured on 3 or 4 occasions . At the end of this time, a small segment of each colony (including tissue and skeleton) was removed by rock saw and preserved for future analysis. This reduced the surface areas of each colony by * 25%. Seawater temperatures were then reduced to 25°C over a period of 4 weeks and then acclimated at this temperature for another month during which time the sawn edges of the coral colonies were overgrown by coral tissue. Finally, corals were restained with Alizarin Red S, physiological rates were measured over a 5-week period as before, and the corals were then sacrificed. Further details on the coral culturing methods are provided in the supplementary information. It is likely that the import procedures and sawing of colonies resulted in some stress to the corals at the start of this study, but it is unlikely that this impacts the data presented here. Corals were cultured for a total of 11 months during this experiment, displayed no evidence of bleaching and maintained photosynthesis, respiration and calcification rates comparable to those of massive Porites spp. in the field (see supplementary information).

Seawater characterisation
The physical characteristics of the culture seawater are detailed in Cole et al. (2018) and are summarised in Table 1. Dissolved nutrients were measured in filtered samples from each reservoir using a flow cell spectrophotometer (Lachat 8000) at the Scottish Association of Marine Science, the UK. [NO 3 -? NO 2 -] and [PO 4 -] fell to low levels (0.04 and 0.00 lM, respectively) in the first week of the 5-week experimental period in the 400 latm pCO 2 treatment at 28°C (Table 1) and the polyps of these corals appeared retracted. After this experimental period, the seawater in this reservoir was discarded and replaced with seawater sourced from the remaining three reservoirs which was bubbled to bring it to seawater pCO 2 of 400 latm before use. The geochemistry of these specimens was only analysed in the skeleton deposited at 25°C. Dissolved nutrients in the remaining treatments did not vary significantly between treatments and were comparable to natural reefs (Szmant 2002).
Seawater samples were collected 2-4 times during each experimental period and analysed by quadrupole ICP-MS (Thermo Scientific X Series) for Mg and Ca. Samples were diluted 1000-fold in 5% HNO 3 (with 5 ppb In as an internal standard) and calibrated against matrix-matched synthetic standards prepared from 1000 lg ml -1 single-element stock solutions (Inorganic Ventures) in 5% HNO 3 . Replicate analyses of IAPSO standard seawater yielded Mg and Ca precision (1r) of 1.4% and 0.6%, respectively.
Seawaters were sampled for d 11 B in the first and last weeks of each 5-week experimental period (with the exception of the 400 and 750 latm treatments at 25°C which were sampled in the first week only). Boron was separated from the seawater matrix by column chromatography (using the boron-specific anionic exchange resin Amberlite IRA 743) and analysed on a Neptune Plus MC-ICPMS (Foster 2008;Rae et al. 2011;Foster et al. 2013). Long-term reproducibility for samples is conservatively estimated at 0.23 permil (2SD). Triplicate analyses of an in-house seawater standard prepared alongside these samples gave a mean value of 39.50 ± 0.09 % (2r).

Estimation of pH ECM
After sacrifice, the coral heads were submerged in 3-4% sodium hypochlorite for C 24 h with intermittent agitation to remove organic contamination, rinsed repeatedly in distilled water and dried. Colonies were sawn perpendicular to the growth surface to produce strips along the maximum growth axes and fixed in 25 mm epoxy resin blocks (Epofix, Struers Ltd.). Blocks were polished using silicon carbide papers (up to 4000 grade, lubricated with water) and polishing alumina (0.05 lm, suspended in water).
The skeletal d 11 B of each experimental period was determined by SIMS using a Cameca 1270 in the School of GeoSciences at the University of Edinburgh. Sections were gold coated and analysed with a primary 16 Obeam of * 7 nA, accelerated at 22 keV and focussed to an oval * 25 9 35 lm. Multiple SIMS analyses were evenly spaced across the skeleton deposited in 2-5 different corallites of each colony over each experimental period. Instrument conditions were energy offset = 0 eV (100 eV window), imaged field = 25 lm, entrance slits 150 lm and exit slits 500 lm (mass resolution was * 2400). Secondary ions were collected by a single electron multiplier cycling the magnetic field through the mass range. Singly charged cations were collected at masses 10 B (11 s per cycle) and 11 B (3 s) yielding typical count rates of * 1300 and 5000 cps, respectively. Each analysis is the sum of 60 cycles. A pre-analysis sputter time of 30 s in spot mode was used to remove surface contamination. Internal reproducibility (the precision at a single point) was calculated from the standard deviation (r) of the 60 cycles in each coral analysis (r/(H60)) and was typically 1.6%. The precision (1r) of multiple analyses (n = 9-25) of each coral sample was typically 1.5% (d 11 B).
All analyses were normalized to a Porites spp. coral standard (d 11 B = 24.8%, B/Ca = 0.364 mmol mol -1 , (Kasemann et al. 2009)). Minor differences in the d 11 B of the SIMS coral standard and the chips of the same coral characterised by bulk methods may affect the accuracy but not the precision of SIMS estimates. A Desmophyllum spp. cold water coral chip (d 11 B & 16.7 ± 1.2% (1r); B/Ca & 0.15 mmol mol -1 ± 3% (1r), based on comparison with the coral standard) which exhibited limited heterogeneity in d 11 B was analysed to confirm that there was no instrumental drift within and between days. Multiple analyses were completed each day (n = 9-23) to yield a 95% using a B (= 1.0266, the mean of two empirical estimates Klochko et al. 2006;Nir et al. 2015), pK B calculated from measured temperatures and salinity and assuming that the d 11 B of the calcification fluid (d 11 B ECM ) is the same as culture seawater (Table 1).

Results
The culture system parameters, seawater and skeletal d 11 B data and calculated pH ECM are summarised in Table 1 Fig. 1). Significant variations in pH ECM occur both between the same coral genotype in different seawater pCO 2 treatments and between different coral genotypes within the same seawater pCO 2 treatment (one-way ANOVA, Table 2). pH ECM is significantly higher in all coral genotypes at seawater pCO 2 180 latm compared to 750 latm at both 25 and 28°C (with the exception of G4 at 25°C, Table 2). Mean pH ECM is significantly lower at 28°C compared to 25°C in corals cultured at 180 latm (paired 2 tailed t test of all 4 coral genotypes, p = 0.00053) but not in the same coral genotypes cultured at the two temperatures at 750 latm (p = 0.65). Minor changes in pHsw occur between 25 and 28°C within some pCO 2 treatments (Table 1). To account for this, we also tested if the proton concentration gradient between the  (Mehrbach et al. 1973), refit by (Dickson and Millero 1987) and for KHSO 4 from (Dickson 1990)  (Fig. 2). Calcification rate and the [H ? ] gradient are significantly correlated at 25°C in genotypes G3 and G6, but all other relationships are not significant (Table 3). This partly reflects the small numbers in the regression analysis (n = 3 or 4), but it is noteworthy that calcification rate and the [H ? ] gradient are inversely related in all corals at 25°C but demonstrate much more mixed behaviour at 28°C.
Only one reservoir was used for each pCO 2 treatment, and to estimate variations between replicate reservoirs, we compare the results from this study to those of a previous study using the same aquarium system and culturing corals

Discussion
Increasing seawater temperature from 25 to 28°C significantly reduced pH ECM (and the H ? gradient between the calcification media and the ambient seawater) at 180 latm but had no significant effect at 750 latm. Coral calcification rates did not change significantly between 25 and 28°C in the 180 latm treatment but increased significantly at high seawater pCO 2 . Previous studies linking temperature and pH ECM generate mixed reports. Temperature had little effect on pH ECM of Acropora spp. cultured at ambient and high pCO 2 (Reynaud et al. 2004;Dissard et al. 2012). pH ECM showed little change in massive Porites spp cultured in an outdoor flume system in response to a 3°C seasonal seawater temperature change at ambient pCO 2 and a small reduction (in pH ECM ) during the cooler months at high pCO 2 (Comeau et al. 2019). In contrast, a broad inverse correlation was reported between temperature and pH ECM in modern field corals (i.e. growing at atmospheric pCO 2 ) from varying localities in Western Australia (Ross et al. 2019). These discrepancies could reflect the role of other influences on pH ECM . Increasing seawater pCO 2 reduced pH ECM more in the dark than the light in some coral species ) and seasonal variations in light availability in the flume experiments may affect pH ECM estimates which combine skeleton deposited during both the day and night (Comeau et al. 2019). The corals in this study were maintained at a constant light regime throughout the study, so this factor is not responsible for changes in pH ECM here. Aragonite precipitation rates are influenced by fluid saturation state (a measure of the relative concentrations of the ions utilised in precipitation), temperature and the presence of biomolecules (Burton and Walter 1987;Kellock et al. 2020). We consider the influences of these factors in the two sections below. 2-] assuming that B(OH) 4 competes with either CO 3 2or HCO 3 for substitution into CO 3 2sites in the aragonite lattice (Allison et al. 2014). Alternatively, the full width half maxima (FWHM) of the v 1 peak in the aragonite Raman spectrum has been postulated as an indicator of seawater/ECM saturation state (De Carlo et al. 2017). Both of these approaches offer promise but we believe that neither of them has yet been constrained accurately enough to be applied to biogenic carbonates. For both approaches, the relationships between the proxy and the seawater saturation state or [CO 3 2-] have been developed from a series of inorganic aragonite precipitation experiments (Holcomb et al. 2016). The Ca 2? consumed during precipitation of the aragonite in these experiments was not replaced and seawater [Ca 2? ] varied significantly within each precipitation (by * 9 3). This resulted in large changes in fluid saturation state within each precipitation and is likely to have significantly affected both aragonite B/Ca (as observed in calcite, Gabitov et al. 2014) and the Raman spectra (De Carlo et al. 2017). This undermines the effective use of these proxies to reconstruct ECM [CO 3 2-] and X. We also note that the inclusion of organic molecules in CaCO 3 influences the FWHM of the v 1 peak (Ihli et al. 2019). There are large variations in the biomolecule compositions of coral skeletons (Coronado et al. 2019;Kellock et al. 2020), and this will complicate any interpretation of ECM X from Raman spectra.
To explore the potential relationship between temperature and coral calcification rate, we use synthetic aragonite precipitation rates observed at 5, 25 and 37°C (Burton and Walter 1987) to calculate the precipitation rates of inorganic aragonite as a function of temperature and X (Fig. 3a,b). We overlay the observed coral calcification rates ) onto this graph assuming that the corals cultured at 400 latm and 25°C have a mean X ECM = 12 (equivalent to that calculated from pH and [CO 3 2-] measured by microsensors in Stylophora pistillata cultured at 400 latm and 25°C (Sevilgen et al. 2019) an arbitrary precipitation rate equivalent to that observed at this X and 25°C in inorganic aragonites, i.e. 2.1 lmol m -2 h -1 (Burton and Walter 1987). We scale linearly all other coral precipitation rates to this value and overlay them onto the graph to estimate X ECM (Fig. 3b). The effects of increasing temperature on coral calcification rate are consistent with decreases in X ECM at low seawater pCO 2 and an increase in X ECM at high seawater pCO 2 . Although inferred X ECM may decline at higher temperatures in the 180 and 260 latm treatments, the impact of temperature on aragonite precipitation is sufficient to offset this reduction. In contrast, at high seawater pCO 2 , the inferred X ECM increases in tandem with temperature enabling the coral to accelerate calcification to a much greater extent than observed at low seawater pCO 2 . Increasing temperature marginally shifts the DIC equilibrium to favour the speciation of CO 3 2at constant pH and DIC (Mehrbach et al. 1973) and marginally increases X at constant pH and [CO 3 2-]. However, the relatively large variations in X ECM suggested by this inorganic aragonite exercise are consistent with a decrease in [DIC] ECM at low seawater pCO 2 and an increase in [DIC] ECM at high seawater pCO 2 .

The role of the skeletal organic matrix in aragonite precipitation
In the above exercise, we assume that coral calcification is only influenced by X ECM and temperature. However, this is an oversimplification. Corals produce a skeletal organic matrix composed of proteins, sugars, polysaccharides and lipids which controls mineral precipitation (Falini et al. 2015). Observations of modern day Porites spp. corals, growing over a range of seawater X, indicate that skeletal density is reduced at low seawater X but that linear extension is not affected (Mollica et al. 2018). These variations are not driven by ECM DIC conditions and likely reflect the role of the skeletal organic matrix in skeletogenesis. Aspartic acid, the most common amino acid in the coral skeletal organic matrix (Cuif et al. 1999), inhibits aragonite precipitation at the concentrations inferred to occur at the coral calcification site (Kellock et al. 2020). The degree of inhibition increases at low seawater X and at higher [aspartic acid]. Skeletal aspartic acid and total amino acid concentrations in cultured Porites spp. corals are positively correlated with seawater pCO 2 (Kellock et al. 2020). Reducing X ECM and increasing skeletal [aspartic acid], as can occur in corals at high pCO 2 , both act to increase the degree of aspartic acid inhibition of aragonite precipitation and can contribute to reduced coral calcification rates under ocean acidification. We know of no reports that constrain the effect of temperature on the skeletal organic matrix, however, changes in the concentration, composition and behaviour of the skeletal organic matrix will influence coral calcification rate. Increasing temperature had no significant effect on pH ECM at high pCO 2 but reduced pH ECM at low seawater pCO 2 . If reduced pH ECM is associated with reduced X ECM , then this offset between pCO 2 treatments may explain why calcification rates remain similar between 25 and 28°C at low seawater pCO 2 (reduced X ECM enhances any aspartic acid inhibition of aragonite precipitation) but accelerates at high seawater pCO 2 (increased X ECM decreases any aspartic acid inhibition of aragonite precipitation).

Insights into controls on pH ECM
Changes in pH ECM between corals and temperatures could reflect changes in calcification (reducing pH ECM ), in the influx/efflux of relevant solutes to/from the calcification site or in the activity of carbonic anhydrase (CA) enzymes that catalyse the interconversion of CO 2 ? H 2 O and HCO 3 -? H ? . Both calcium adenosine triphosphatase (Ca-ATPase) and a bicarbonate anion transporter (BAT) have been localised to the plasma membrane of the  Fig. 3 Temperature effects on precipitation rates in corals and inorganic aragonites. a Inorganic aragonite precipitation rates at three temperatures as a function of X (aragonite saturation state, redrawn from (Burton and Walter 1987)). b Mean coral precipitation (calcification) rates in each seawater pCO 2 treatment ) overlaid onto inorganic aragonite precipitation rate contours (in pale grey for different X) replotted as a function of temperature. We assume that corals cultured at 400 latm and 25°C have a X ECM = 12 (Sevilgen et al. 2019) and a precipitation rate equivalent to that observed at X = 12 and 25°C in inorganic aragonites (Burton and Walter 1987). We scale linearly all other coral precipitation rates to this value calicoblastic cells and remove H ? and add HCO 3 to the ECM, respectively (Zoccola et al. 2004(Zoccola et al. , 2015, see Fig. 4. The activities of calcification-related enzymes and solute carriers and the expression of the genes which encode them can be affected by both temperature and seawater pCO 2 (Edge et al. 2005;Barshis et al. 2013;Kenkel et al. 2013;Maor-Landaw et al. 2014;Zoccola et al. 2016;Kurihara et al. 2018;Yuan et al. 2018). Temperature increases, below the denaturing threshold, usually promote enzyme activity, e.g. increasing temperature enhances the efficiencies of coral Ca-ATPase (Ip et al. 1991) and coral CAs (Zoccola et al. 2016) over the temperature range studied here (25-28°C). Ca-ATPase uses ATP as an energy source for Ca 2? transport; however, respiration rates did not change significantly between 25 and 28°C in the corals cultured at 750 latm pCO 2  implying that the corals are maintaining a steady energy expenditure (Fig. 4). Ca-ATPase is only one component of the energy budget and temperature-driven increases in the activities of other biological processes (reducing their energy demand) could offset any increase in energy demand by Ca-ATPase. In contrast, respiration rates in corals cultured at 180 latm pCO 2 were significantly lower (two-tailed paired t test, p = 0.03) at 28°C (by * 20%) than at 25°C . This implies that coral metabolic costs, including calcification, are reduced at the higher temperature although individuals maintain similar calcification rates at 25 and 28°C in spite of this. In this case, the reduction in energy expenditure, e.g. reducing synthesis of calcificationbased enzymes and carriers, could be balanced by temperature enhancement of the activities of the enzyme and solute carriers which are present coupled with the temperature promotion of aragonite precipitation. This hypothesis is consistent with evidence that temperaturedriven increases in enzyme activity may even offset gene downregulation in response to ocean acidification (Zoccola et al. 2016).
Seawater acidification increases the permeability of the coral tissue and could accelerate the paracellular transport of ions (H ? , DIC species, Ca 2? ) along concentration gradients to and from the calcification site (Venn et al. 2020). Increasing temperature in that study had no effect on paracellular permeability. The [H ? ] sw -[H ? ] ECM gradient in the current study is much higher at low seawater pCO 2 (Venn et al. 2013Schoepf et al. 2017)   . In this study, corals exhibit similar inverse correlations at 25°C but more mixed relationships at 28°C (Fig. 2) . 1 and 2). Further resolution of the relationships between temperature, calcification and metabolism requires observations over a greater range of temperatures (as in Coles and Jokiel 1977).

Implications for coral reefs under future climate change
This study was conducted under laboratory conditions, and there were no diurnal variations in light availability and seawater DIC chemistry as observed on natural reefs (Dai et al. 2009). Furthermore, we do not know the typical seawater pH of the reef site in which the corals grew and were acclimated before collection. However, our study yields insights into the likely interactions of seawater pH and temperature in massive Porites spp. corals. At 750 latm pCO 2 , calcification rates at 28°C exceed those at 25°C  suggesting that temperature increases, below the thermal stress threshold, can at least partially offset the impact of ocean acidification. Future increases in seawater temperatures may offset ocean acidification effects in individuals of this genus living below their upper thermal stress limits. However, elevating seawater temperature beyond the stress threshold reduced both calcification and pH ECM in the branching corals Stylophora pistillata and Pocillopora damicornis over a wide range of seawater pH (Guillermic et al. 2021). Thermally induced bleaching may also decrease pH ECM (Dishon et al. 2015 (Reynaud et al. 2004;Dissard et al. 2012) although increasing temperature at high seawater pCO 2 acted to significantly decrease calcification in Acropora intermedia (Anthony et al. 2008). While some coral species may acclimatise or adapt to climate change (IPCC 2019), seawater temperature increases are unlikely to benefit the majority of corals which are already living close to their maximum thermal tolerance limits. Coral reefs are predicted to decline greatly at even a 2°C temperature increase compared to pre-industrial levels (IPCC 2019).
Acknowledgements This work was supported by the UK Natural Environment Research Council (award NE/I022973/1) to AAF and NA. NERC Scientific Services provided access to the ion microprobe, and we are indebted to Richard Hinton and John Craven (EIMF, University of Edinburgh) for their assistance with the analyses. We thank Dave Steven, Mark Robertson, Casey Perry, Mike Scaboo and Andy Mackie for their assistance with the culture system build. We thank two anonymous reviewers for constructive comments on the first draft of this manuscript.

Declarations
Conflict of interest On behalf of all authors, the corresponding author states that there is no conflict of interest.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons. org/licenses/by/4.0/.