The geologic history of seawater pH

See allHide authors and affiliations

Science  10 Mar 2017:
Vol. 355, Issue 6329, pp. 1069-1071
DOI: 10.1126/science.aal4151

The long view of ocean pH

The acid-base balance of the oceans has been critical in maintaining Earth's habitability and allowing the emergence of early life. Despite this importance, systematic estimates of historical seawater pH are lacking. Halevy and Bachan developed a model of seawater chemistry and pH over time scales exceeding ∼100 million years. Their highly robust probabilistic history of seawater pH and chemistry reflects evolving properties of Earth's atmosphere, oceans, and crust. Seawater pH increased from early Archean values of ∼6.5 to 7.0 to more recent values of ∼7.5 to 9.0 mostly as a result of solar brightening and decreasing interaction between seawater and oceanic crust.

Science, this issue p. 1069


Although pH is a fundamental property of Earth’s oceans, critical to our understanding of seawater biogeochemistry, its long-timescale geologic history is poorly constrained. We constrain seawater pH through time by accounting for the cycles of the major components of seawater. We infer an increase from early Archean pH values between ~6.5 and 7.0 to Phanerozoic values between ~7.5 and 9.0, which was caused by a gradual decrease in atmospheric pCO2 in response to solar brightening, alongside a decrease in hydrothermal exchange between seawater and the ocean crust. A lower pH in Earth’s early oceans likely affected the kinetics of chemical reactions associated with the origin of life, the energetics of early metabolisms, and climate through the partitioning of CO2 between the oceans and atmosphere.

A key parameter of ocean chemistry, pH ties together climate, redox, and biological activity. Biosynthesis of adenosine triphosphate (ATP) in all living organisms exploits cross-membrane proton gradients (1), and ambient pH affects the saturation state of carbonate minerals, including biogenic calcium carbonates (2). Moreover, as the stability of both primary crustal and authigenic minerals depends on pH (3), so do the sources and sinks of major and trace elements from the dissolution and precipitation of such minerals. Solution pH further governs the speciation, solubility, and sorption tendency of many metals, some of which are bioessential nutrients (e.g., Fe and Zn), and others that serve as proxies for Earth’s surface oxidation state (e.g., Mo, Cr, and U) (4). Finally, for a given carbon pool size in the ocean-atmosphere, seawater pH governs the partitioning of carbon between dissolved inorganic carbon (DIC) and atmospheric CO2, the latter of which is a dominant factor in Earth’s climate. Hence, the evolution of seawater pH has been important for both the organic and inorganic sinks in the long-term carbon cycle, for the attendant oxygen cycle, for the cycles of all major seawater ions, and for the emergence of biocalcification.

Despite its importance, the pH of the oceans over most of Earth history remains poorly constrained. The long timescale of interest and the preservation state of Precambrian carbonates preclude the use of direct proxies, such as boron isotopes, to constrain seawater pH evolution (5). Independent constraints on the atmospheric partial pressure of CO2 (pCO2) and seawater calcium concentrations ([Ca2+]), together with an assumption of CaCO3 saturation, allow estimation of the pH of surface waters [e.g., (6)]. In the absence of robust information on Precambrian pCO2 and [Ca2+], constraints on [HCO3]/[Ca2+] from the geologic record of sulfate evaporites have yielded wide pH range estimates as a function of pCO2 (79). More accurate predictions are acknowledged to require a more sophisticated model of the relevant geochemical cycles (8). Finally, with assumptions about the relative importance of low-temperature interaction with alkaline rocks (a source of alkalinity) and high-temperature hydrothermal alteration of oceanic crust (a source of acidity), but with no explicit attempt to quantify pH, theoretical studies have concluded that the early ocean was alkaline (10), acidic (11), or approximately neutral (12).

To constrain the evolving marine acid-base balance, we developed a statistical model of seawater pH as a function of pCO2 and the suite of parameters that govern ocean chemistry (13). Specifically, we solve the charge balance of the major seawater ions, [HCO3] + 2[CO32–] + [HS] – [NH4+] + [OH] – [H+] = [Na+] – [Cl] + [K+] + 2[Mg2+] + 2[Ca2+] + 2[Fe2+] – 2[SO42–](1)where conservative ions (strong acids and bases) and weak acids and bases are on the right and left of Eq. 1, respectively. We relate [OH], [HCO3], and [CO32–] directly to [H+] and pCO2 through the reactions of water and DIC speciation. An assumption of CaCO3 saturation (or some degree of supersaturation, as in the present ocean) relates [Ca2+] to [H+] and pCO2. We explicitly model [Na+], [Cl], [K+], and [Mg2+] as balances of their pH- and pCO2-dependent sources and sinks, and we assign [HS], [NH4+], [Fe2+], and [SO42–] to within estimated ranges for different times in Earth history. With prescribed or modeled ion concentrations, Eq. 1 becomes a quartic equation in [H+], which we solve numerically to give seawater pH as a function of pCO2. We draw model parameters from wide distributions that represent the uncertainty in their values (13) to provide probabilities for the time-dependent chemical composition and pH of seawater.

With present-day concentrations of the major seawater ions and preindustrial pCO2 (280 ppm), the calculated pH is 8.17, close to the preindustrial average pH of surface waters (14). Forced with Phanerozoic proxy and model estimates of pCO2 and with seawater [SO42–] ranges from halite fluid inclusions (13), Phanerozoic seawater pH varies between likely values of ~7.5 and ~9.0 (Fig. 1A), irrespective of whether [Ca2+], [Mg2+], and [K+] are constrained by halite fluid inclusions or modeled (fig. S1). This suggests that, although factors that affect seawater ion concentrations over timescales shorter than a few 100 million years (e.g., tectonic cycles, paleogeography, weathered lithology, and climate variability) are not explicitly accounted for, the uncertainty envelope on model pH values brackets their likely effects. In other words, although the most probable model pH cannot be regarded as an exact history of seawater pH, the true value has likely been within the uncertainty envelope. Over Earth history, the model ocean remains unsaturated with sulfate evaporites, and the marine DIC pool never exceeds the current surface carbon reservoir (ocean + atmosphere + sedimentary rocks ≈ 1022 g) (8), a conservative estimate for the total carbon fluxed into the ocean-atmosphere over time.

Fig. 1 The long-term evolution of seawater pH.

(A) Results of ~7 × 105 default model simulations (frequency in color contours and 95% of values within gray line). (B) Comparison of full model results with results obtained under constant pCO2 = 280 ppm, and under constant pCO2 = 280 ppm and constant modern heat flux (95% of values within gray field, and light orange and dark orange lines, respectively), demonstrating the control of pCO2 and heat flux on the evolution of seawater pH. (C) Comparison of full model results with results obtained with nW = 0.0 and nW = 0.4 (95% of values within gray field and light orange and dark orange lines, respectively), demonstrating the effect of nW on model uncertainty. Ma, million years ago.

The model suggests that over Earth history, seawater pH has increased from between ~6.5 and 7.0 in the early Archean to between ~7.5 and 9.0 over Phanerozoic time (Fig. 1A). Several factors are responsible for this secular increase, most notably a pCO2 decrease in response to solar brightening (15). For given concentrations of the conservative ions in seawater, higher pCO2 leads to lower pH. However, if the intensification of weathering rates and the associated delivery of weathering-derived cations to the oceans is large enough with increasing pCO2, then it may partially compensate for the acidification. With wide parameter ranges, we account for these effects and find that the effects of increasing pCO2 are incompletely compensated for by increased weathering fluxes, and that pH decreases. When pCO2 is held constant at 280 ppm, early Archean pH is ~7.5 to 8.0 (instead of ~6.5 to 7.0) (Fig. 1B). The remaining decrease of ~1 pH unit is due to Earth’s higher early heat flux, which leads to more rapid production of ocean crust and larger seawater circulation fluxes through this crust (12, 16). As both near-axis and off-axis hydrothermal circulation is a sink for Mg2+ (17), a higher heat flux leads to lower [Mg2+] (fig. S2D). The decreasing importance of Mg2+ in the marine charge balance leads to an increase in [H+] (i.e., a pH decrease). When both pCO2 and Earth’s heat flux are held constant, no secular pH trend remains (Fig. 1B).

As the geologic drivers of the increasing pH trend (solar brightening and mantle cooling) are largely undisputed, the sign of the pH trend over Earth history is robust, even if its magnitude is uncertain. Two potentially important factors are unaccounted for in the model, continental growth and evolving crustal lithology, both of which could affect the pH trend, but in opposing directions. Smaller early continental area implies lesser continental alkalinity sources (lower pH), but more mafic crystalline rocks (10) and less chemically mature sediments (18) imply greater release of alkalinity per unit mass of weathered crust (higher pH). The net effect on seawater pH is uncertain but would have been important only before the late Archean and likely of secondary importance to the long-term trend we present.

The width of the pH envelope reflects uncertainty in the values of only a few model parameters—most important, the abundances of the acids in the ocean (pCO2 and seawater [SO42–]) and the dependence of weathering rates on pCO2. Operation of the silicate-weathering negative climate feedback, as assumed here, means that CO2 levels adjust to the value required for temperature-dependent continental silicate weathering rates (15) and seafloor alteration rates (19) to balance volcanic outgassing of CO2. On the timescales of interest here, this implies a near-constant temperature, and a dependence of weathering rates on pCO2, mainly through their dependence on rainwater and groundwater pH. With increasing pCO2, more acidic rainwater could result in more rapid chemical weathering reactions. We quantify the weathering enhancement factor due to rainwater pH asEmbedded Image(2)where the proton concentration in rainwater [H+]rain ≈ (pCO2 × KH × K1)1/2, KH is the Henry's law constant for CO2, K1 is the carbonic acid first dissociation constant, Embedded Image is the proton concentration in preindustrial rainwater exposed to 280 ppm CO2, and nW is an exponent that describes the dependence of weathering rates on rainwater pH. Among other factors, nW embodies the effects of pH buffering by soils and of topography (20) on the sensitivity of weathering rates to pCO2. To account for variability and uncertainty in its value in natural weathering environments, we draw nW from a distribution with a peak at 0.2 and with ~95% of values between 0 and 0.5 (13). At the higher end of this range, a pCO2 increase leads to increased delivery of alkalinity to the oceans, and the pH decrease is modulated. At the lower end, continental weathering is unaffected by the change in rainwater pH, and seawater pH decreases more strongly with increasing pCO2. These two end-member cases cover the range of variability in model seawater pH (Fig. 1C), and remaining scatter at a given value of nW is mainly due to poor constraints on pCO2 and seawater [SO42–] (fig. S3). An improvement in available pCO2 and seawater [SO42–] estimates and, importantly, in understanding the dependence of weathering rates on pCO2 (nW) would reduce uncertainty in early seawater pH.

The pH uncertainty envelope widens with decreasing age (Fig. 1A) because of a shift over Earth history in the relative importance of continental and seafloor weathering. A high early heat flux and associated large hydrothermal water fluxes emphasize the importance of seafloor weathering, whereas continental weathering (and the uncertainty in its fluxes due to the uncertain value of nW) increases in relative importance as Earth ages and the heat flux decreases. Any continental growth scenario would further stress the early role of seafloor alteration and the later importance of continental weathering.

Several abrupt shifts in pH (Fig. 1A) are due to prescribed changes in pCO2, in response to presumed changes in non-CO2 components of the atmospheric greenhouse (e.g., methane) (13). The presence of non-CO2 greenhouse agents decreases the pCO2 required to balance CO2 outgassing by silicate weathering, which is implied by the silicate-weathering feedback. The appearance of methanogens by 3.5 billion years ago (Ga) (21) would have increased methane fluxes into the anoxic atmosphere and allowed the accumulation of hundreds to a few thousands ppm methane (22). An associated halving of pCO2 results in the pH increase between 3.5 and 4.0 Ga (13). The pH decrease in the early Proterozoic reflects a suggested decrease in reduced greenhouse gas concentrations due to the rise of atmospheric oxygen, and the associated increase in pCO2 (13). Although the exact timing of these pH shifts is uncertain and their magnitude often small (a few 10ths of a pH unit), the qualitative behavior is well understood.

The concentrations of oxygen-sensitive ionic seawater species are likely to have changed over Earth history, too. Within reasonable estimates of [Fe2+], [HS], and [NH4+] in the early anoxic ocean, their effect on the charge balance is negligible. In contrast, sulfate contributes importantly to the present-day charge balance and has likely done so for the entire Phanerozoic and large parts of the Proterozoic. Archean [SO42–] was too low (13) to have affected the seawater charge balance. However, after the rise of atmospheric oxygen, [SO42–] increased, perhaps to between several hundred μM and a few mM (23). Over this range of concentrations, sulfate begins to affect the charge balance and contributes both to the pH decrease following the rise of oxygen (Fig. 1A) and to the growing uncertainty in model seawater pH values from the Archean into the Proterozoic.

Our results bear on several topics of ongoing research. The probable [HCO3]/[Ca2+] ratio is below 2 for the majority of Earth history (fig. S2I), which suggests that the near-absence of sulfate evaporites before ~1.8 Ga is not due to depletion of Ca2+ during seawater evaporation (e.g., 8) but due to low seawater [SO42–]. These findings are in agreement with recent estimates of seawater carbonate chemistry from calcium isotopes in Neoarchean and Paleoproterozoic sulfate evaporites, which suggest relatively low [HCO3]/[Ca2+] ratios (9). In the early and late Proterozoic, when evidence for repeated glaciation of near-global extent independently suggests low background pCO2 (24, 25), pH is expected to be high within its uncertainty envelope. Formation of primary and early diagenetic talc in Neoproterozoic carbonate successions required porewater pH ≥ ~8.6 (26, 27), consistent with the upper boundary of our simulated pH at that time (~8.8). We suggest that low Neoproterozoic pCO2 resulted in relatively high pH, which enabled talc precipitation in the water column and shallow sediments. We are unaware of observational evidence for relatively high pH surrounding Paleoproterozoic glaciations but offer such observations as tests of the model proposed here. Variability in transition metal concentrations or isotopic compositions is interpreted to reflect changes in oxidation state (4), but the speciation of these metals depends on pH (28) within the range relevant to Earth history. Do some of the transitions or spikes in the records of these metals reflect a change in seawater pH rather than its oxidation state? Finally, the pH dependence of oxygen isotope fractionation between aqueous solutions and carbonate minerals (29) implies that early Archean carbonates be enriched in 18O by ~2 to 3‰ relative to carbonates formed at pH 8.0 to 8.5. Thus, the long-term increase in 18O/16O of marine carbonates [~15‰ over ~3 Ga (30)] may underestimate the difference between modern and Archean seawater temperature or 18O/16O, the two main explanations for the secular trend.

Correction (15 March 2017): Report: "The geologic history of seawater pH" by I. Halevy and A. Bachan (10 March 2017, p. 1069). In Equation 1, the operation signs and the order of terms on the left side of the equation were incorrect. The rest of the text, including the description of Eq. 1, is correct as it appeared in the original version. The results of the study remain unchanged.

Supplementary Materials

Materials and Methods

Supplementary Text

Figs. S1 to S15

Tables S1 to S3

References (3193)

References and Notes

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.Materials and methods are available as supplementary materials.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
  56. 56.
  57. 57.
  58. 58.
  59. 59.
  60. 60.
  61. 61.
  62. 62.
  63. 63.
  64. 64.
  65. 65.
  66. 66.
  67. 67.
  68. 68.
  69. 69.
  70. 70.
  71. 71.
  72. 72.
  73. 73.
  74. 74.
  75. 76.
  76. 77.
  77. 78.
  78. 79.
  79. 80.
  80. 81.
  81. 82.
  82. 83.
  83. 84.
  84. 85.
  85. 86.
  86. 87.
  87. 88.
  88. 89.
  89. 90.
  90. 91.
  91. 92.
  92. 93.
  93. 94.
  94. Acknowledgments: I.H. acknowledges funding from the European Research Council Starting Grant 337183. I.H. conceived the research, developed, and analyzed the model, and wrote the paper. A.B. contributed to the discussion and writing.
View Abstract

Stay Connected to Science

Navigate This Article