Evolution of the ocean's “biological pump”

Earth history is punctuated by a huge variety of transitions and perturbations in climate and global biogeochemical cycles. These may be linked to major extinctions or evolutionary innovations, and may exhibit evidence for greenhouse warming and CO₂ release and hence potentially hold direct future-relevant information (1) or may be associated with ice ages. Arguably, no event is more enigmatic or has been more keenly debated than the occurrence of extreme glaciation during the Neoproterozoic (1,000–542 Ma) (2), when, in two separate episodes, the global ocean potentially attained complete sea-ice cover to create a “snowball Earth” (3). One of the main barriers to a full understanding of these intense glacial episodes has been in identifying the trigger; as nothing comparable occurs at any time in the preceding approximately 1,500 Ma or afterward during the Phanerozoic. The mechanism for the initial cooling must also be consistent with a pronounced negative excursion recorded in the carbon isotopic composition (i.e., δ¹³C) of marine carbonates (2). This latter criterion is particularly challenging because, although the oxidation of reduced, biological forms of carbon such as organic matter and hydrate methane are highly depleted in ¹³C and hence are able to drive ocean and atmosphere δ¹³C negative, the release of greenhouse gases such as CO₂ (and CH₄) will tend to prevent, not cause, a deep ice age (4). In PNAS, Tziperman and colleagues (5) square the circle between cooling [and a lower partial pressure of CO₂ (pCO₂) in the atmosphere] and negative-trending δ¹³C by recognizing that alternative pathways for oxidizing organic matter would have been important and that carbon cycling in the ocean may have undergone profound changes around this time.

In today's ocean, the fixation of carbon at the surface by photosynthesizing organisms and subsequent sinking of a sizable fraction (15–20%) of total productivity creates a strong vertical transport that dominates the distribution of carbon, nutrients, and oxygen in the ocean (Fig. 1A), known as the ‘biological pump’. However, the ocean carbon cycle during the Precambrian may have been very different (6), a possibility that Tziperman et al. (5) make use of in their model analysis. They explain the isotopic decline by an increase in the efficiency of organic carbon export to depth that induces a decrease in the reservoir of isotopically depleted dissolved organic matter (DOM). The authors avoid the seemingly inevitable consequence of oxidizing organic matter—higher atmospheric pCO₂—by first recognizing that, in the poorly oxygenated Neoproterozoic ocean, bacterial metabolism of organic matter would tend to proceed coupled to sulfate (SO₄²⁻) reduction rather than using O₂ (7). Removal of SO₄²⁻ has the effect of making the ocean less acidic, lowering atmospheric pCO₂. However, H₂S produced as a byproduct of sulfate reduction would eventually reach the oxygenated surface layers of the ocean and sulfate concentrations regenerated. Tziperman et al. (5) get around this by noting that the H₂S can be removed by reaction with dissolved Fe and buried as the mineral pyrite (FeS₂)—they additionally suggest that carbon could be buried as siderite (FeCO₃)—leaving ocean pH higher and pCO₂ lower.

Fig. 1.

Illustration of potential differences in modern vs. ancient carbon cycling. (A) Illustration of the operation modern biological pump (12) shows it is dominated by a strong vertical settling flux of particulate organic matter that decreases approximately exponentially with depth. Some of this degrading particulate flux is converted into refractory DOM (black arrows) and persists in the ocean. A large flux of DOM is produced in the surface ocean, but is predominantly highly labile (light gray) and semilabile (dark gray) and hence is not transported far. Overall, the ocean reservoir of dissolved inorganic carbon is the dominant source of exchangeable carbon at the Earth's surface. Carbon inventories for the modern (pre-Industrial) ocean are from Sundquist and Visser (15) and are in units of PgC (or GtC). (B) Illustration of how carbon cycling may have differed in the less well oxygenated Precambrian ocean. A fraction of the DOM produced in the modern ocean that was rapidly degraded now becomes refractory (dark gray). Both atmospheric and ocean inorganic carbon reservoirs are likely to have been larger (16), but both together are potentially dwarfed by the size of the ocean DOM reservoir required in the hypothesis of Rothman et al. (6).

This, and analogous explanations for other isotopic excursions (8–10), depend on the initial presence of an ocean inventory of DOM orders of magnitude larger than exists in today's ocean. However, the DOM cycle of the modern ocean is far from well understood (11, 12), particularly with regard to the mechanisms of formation and destruction of the more recalcitrant fractions that dominate the overall inventory. More recalcitrant DOM may, for instance, be derived directly from plankton or be formed during the bacterially mediated breakdown of particulate organic matter. Removal may be by scavenging in the water column or photolysis at the surface (11), but seemingly, not significantly via direct bacterial metabolism (13). This uncertainty makes it all the more difficult to deduce the nature of carbon cycling in the ancient ocean. Did the ocean start out with an ineffective vertical transport and recycling of particulate organic matter, with the biological pump dominated by ocean circulation and a massive reservoir of DOM (Fig. 1B)? Following the earlier theoretical work by Rothman et al. (6), this is the line that Tziperman et al. (5) take.

What has not been adequately explored is what the characteristics of an ancient DOM reservoir (if it existed) were and how it came to build up in the first place. Recent analysis of the energetic potential from degrading different organic matter fractions offers a clue—a range of molecular structures such as membrane-type compounds that are degraded by bacteria in a well oxygenated ocean may have been effectively recalcitrant in the Precambrian ocean (14). The accumulation of these compounds could potentially create a massive DOM reservoir. If so, expanding anoxic and sulfate-reducing zones and increasing the transport of organic matter into them might be expected to leave a greater range and total inventory of organic compounds undegraded. This is seemly at odds with the shrinking DOM reservoir model of Tziperman et al. (5) although an increased particulate flux would tend to increase the removal of DOM by scavenging. One further complication here is that, under certain conditions, oxidation reactions with goethite (α-FeOOH) can become more favorable than with sulfate (14). Evidence for the occurrence of a Fe-rich and S-poor deep ocean associated with Neoproterozoic glaciations then raises the possibility of DOM drawdown by a switch from sulfate reduction to dissimilatory iron reduction (5). However, this would then conflict with the inferences of Swanson-Hysell et al. (10) who envisage deep ocean concentrations of dissolved Fe and DOM both increasing together during the late Neoproterozoic.

The Neoproterozoic is undoubtedly a time of profound change in the degree of ocean oxygenation and in the evolution of marine organisms and ecosystems (8), and represents the transition to the more familiar Earth system of the Phanerozoic. It seems eminently reasonable that equally profound changes in marine carbon cycling, and specifically involving DOM, might occur, which hints at new and fundamental links between evolutionary innovation and global carbon cycling and climate (5). However, to move further forward, we need to better understand the extraordinary range of molecular structures that constitutes organic matter, how different ecosystems may produce different proportions of recalcitrant material, how bacteria transform labile fractions to recalcitrant, and how the selectivity of degradation differs between different redox environments. Improved process-based understanding of the role of DOM in the ocean will aid the elucidation not only of the ancient world but also how marine carbon cycling may respond to future climate change (12).