Reconsidering the role of carbonate ion concentration in 1 calcification by marine organisms 2

Marine organisms precipitate 0.5-2.0 gigaton of carbon as calcium carbonate (CaCO 3 ) every 9 year with a profound impact on global biogeochemical element cycles. Biotic calcification 10 relies on calcium ions (Ca 2+ ) and generally on bicarbonate ions (HCO 3- ) as CaCO 3 substrates 11 and can be inhibited by high proton (H + ) concentrations. The seawater concentration of 12 carbonate ions (CO 32- ) and the CO 32- -dependent CaCO 3 saturation state (Ω CaCO3 ) seem to be 13 irrelevant in this production process. Nevertheless, calcification rates and the success of 14 calcifying organisms in the oceans often correlate surprisingly well with these two carbonate 15 system parameters. This study addresses this dilemma through rearrangement of carbonate 16 system equations which revealed an important proportionality between [CO 32- ] or Ω CaCO3 and 17 the ratio of [HCO 3- ] to [H + ]. Due to this proportionality, calcification rates will always 18 correlate equally well with [HCO 3- ]/[H + ] as with [CO 32- ] or Ω CaCO3 when temperature, 19 salinity, and pressure are constant. Hence, [CO 32- ] and Ω CaCO3 may simply be very good 20 proxies for the control by [HCO 3- ]/[H + ] where [HCO 3- ] would be the inorganic carbon 21 substrate and [H + ] would function as calcification inhibitor. If the “substrate-inhibitor ratio” 22 (i.e. [HCO 3 - ]/[H + ]) rather than [CO 3 2- ] or Ω CaCO3 controls biotic CaCO 3 formation then some 23 of the most common paradigms in ocean acidification research need to be reviewed. For 24 example, the absence of a latitudinal gradient in [HCO

and is a function of temperature, salinity, and pressure (Mucci, 1983 2+ ] is rather constant in seawater (Kleypas et al., 50 1999). 51 Biogenic CaCO 3 is mainly present as calcite or aragonite, which have different crystal 52 structures and solubility. Calcite is predominantly formed by coccolithophores, foraminifera, 53 and some crustaceans while aragonite is typically found in scleractinian corals. Molluscs can 54 have both calcite and aragonite. Echinoderms and octocorals build calcite with a large fraction 55 of magnesiumn (Mg) included in the crystal lattice (Mann, 2001 (Jokiel, 2013;Jokiel et al., 2014). 112 The present study builds up on these previous findings and aims to refine the thought that 113 calcification is not controlled by a single carbonate chemistry parameter but reacts to a 114 combination of two or more. Therefore, attention will be drawn to a potentially important 115 proportionality between [CO 3  (Millero, 2010), K HSO4 by Dickson (1990), and K sp * determined by Mucci (1983  inorganic carbon are highly diverse among the various calcifying taxa so that generalization 222 of physiological principles would be difficult (see section 3.6 for a discussion on this topic). It 223 may therefore be helpful to approach this question differently and ask more generally whether 224 HCO 3 or CO 3 2would be the more suitable inorganic carbon substrate for calcification. Three 225 different perspectives will be addressed in the following. contributes ~90% to the total DIC pool while CO 3 2contributes less than 10%. Thus, 229 molecular CO 3 2transporters would require a nine times higher affinity to their substrate than 230 HCO 3 transporters. It may therefore make more sense for an organism to rely on the largest 231 inorganic carbon pool if molecular transporters take the ions directly from seawater 232 (Mackinder et al., 2010). 233

Homeostasis 234
The hydration timescale of CO 2 (CO 2 + H 2 O ↔ HCO 3 -+ H + ) is comparatively slow (~10 235 seconds), while the hydrolysis of HCO 3 -(HCO 3 -↔ CO 3 2-+ H + ) is fast (~10 -7 seconds; Zeebe 236 and Wolf-Gladrow, 2001; Schulz et al., 2006). Assuming a transcellular pathway, selectively 237 incorporated CO 3 2that is transported through cytosol with a typical pH around ~7.0 -7.4 238 (Madshus, 1988) would quickly turn into HCO 3 unless the transfer is faster than 10 -7 seconds. 239 In the likely case that the transfer takes longer, CO 3 2would bind a proton in the cytosol and 240 be transported as HCO 3 to the site of calcification where the proton would subsequently be 241 released back to the cytosol during CaCO 3 precipitation. Hence, the cytosolic pH would 242 remain stable in case of selective CO 3 2uptake as long as CO 3 2uptake and CaCO 3 243 precipitation occur at the same rate. However, both processes may occasionally run out of 244 equilibrium for short periods. In these cases, the utilization of CO 3 2as inorganic carbon 245 source would constitute a substantial risk for the orgamisms' pH homeostasis. Excess ] is highly variable. In the 260 habitat of a temperate coraline algae, for example, typical diurnal pH fluctuations can range 261 from ~8.4 at day to ~7.6 at night (Cornwall et al., 2013  increases only marginally (Fig. 1,

Similarities and differences between the DIC/[H + ] and the [HCO 3 -]/[H + ] ratio 305
In a series of papers Jokiel (2011aJokiel ( , 2011bJokiel ( , 2013 proposed that carbonate chemistry controls 306 calcification rates in corals through the combined influence of DIC ("reactant") and H + 307 correlation starts to increasingly deviate from linearity. In the oceans, noticeable deviations 316 start in the pCO 2 range below 250-500 µatm, where an exponentially increasing fraction of 317 the DIC pool is present as CO 3 2- (Fig. 3) long as T, S, and P are constant (Fig. 3). 322 Whether the DIC/[H + ] ratio proposed by Jokiel (2011aJokiel ( , 2011bJokiel ( , 2013  ] or Ω CaCO3 as the most influential parameter. Before starting the discussion I would 332 like to emphasize, however, that carbonate chemistry patterns discussed here are just one 333 among other abiotic (e.g. temperature or light) or biotic (e.g. food availability or competition) 334 factors which must also be taken into consideration when trying to understand the patterns of 335 calcification in the oceans. 336  (Fig. 4C) (Fig. 7). 372

Implications for ocean acidification research 373
The ongoing perturbation of the surface ocean by anthropogenic CO 2 causes a decline of 374 Thus, when high latitude organisms find a way to efficiently protect their crystal skeletons 391 from corrosive seawater, then they may not be more vulnerable to ocean acidification than 392 their warm water counterparts. 393

Limitations and uncertainties 394
This study has argued that a substrate-inhibitor ratio like [ Areas with corrosive conditions will expand under ocean acidification (Orr et al., 2005) so 419 that CaCO 3 dissolution becomes more widespread problem for future calcifiers. However, 420 dealing with dissolution of CaCO 3 is only of secondary relevance for living organisms as 421 everything that dissolves needs to be formed in the first place. Hence, although dissolution 422 processes cannot be left unconsidered, it is reasonable from a biological point of view to focus 423 on the processes that control formation of CaCO 3 .

CO 2 as inorganic carbon source for calcification 463
Some organisms receive significant amounts of inorganic carbon used for calcification from 464 respiratory sources (Pearse, 1970;Erez, 1978;Sikes et al., 1981;Tanaka et al., 1986;Furla et 465 al., 2000). Here, organisms do not exclusively rely on direct inorganic carbon utilization from 466 seawater but supplement calcification to a variable degree with CO 2 gained intracellularly 467 from respired biomass. This CO 2 utilization may be further strengthened (1) when metabolic 468 CO 2 is 'trapped' inside the organisms through the establishment of pH gradients which limit 469 the diffusive loss of CO 2 (Bentov et al., 2009, Glas et al., 2012b  processes are physiologically coupled within photoautotrophs (Paasche, 2002;Allemand et 479 al., 2004). Accordingly, calcification rates will be affected indirectly when photosynthesis is 480 CO 2 limited (Bach et al., 2015). A valuable measure to determine the potential of CO 2 to limit 481 growth and photosynthesis is K 1/2 which denotes the CO 2 concentration where the process 482 runs at half of its maximum. Available K 1/2 measurements suggest that CO 2 limitation mostly 483 occurs well below CO 2 concentrations typically encountered by the organisms in their 484 respective habitats (Rost et  Allemand, D., Ferrier-Pagès, C., Furla, P., Houlbrèque, F., Puverel, S., Reynaud, S., 524 Tambutté, É., Tambutté