Glacial deep ocean deoxygenation driven by biologically mediated air–sea disequilibrium

Abstract

Deep ocean deoxygenation inferred from proxies has been used to support the hypothesis that lower atmospheric carbon dioxide during glacial times was due to an increase in the strength of the ocean's biological pump. This relies on the assumption that surface ocean oxygen (O 2 ) is equilibrated with the atmosphere such that any O 2 deficiency observed in deep waters is a result of organic matter respiration consuming O 2 and producing dissolved inorganic carbon. However, this assumption has been shown to be imperfect because of disequilibrium. Here we use an Earth System Model tuned to a suite of observations, which reproduces the pattern of glacial-to-Holocene oxygenation change seen in proxy data, to show that disequilibrium plays an important role in glacial deep ocean deoxygenation. Using a novel decomposition method to track O 2 , we find a whole-ocean loss of 33 Pmol O 2 from the preindustrial to the Last Glacial Maximum (LGM) despite a 27 Pmol gain from increased solubility due to cooler temperatures. This loss is driven by biologically-mediated O 2 disequilibrium, which increases from contributing 10% of the reduction of the O 2 inventory from solubility equilibrium in the preindustrial, to 27% during the LGM. Sea

Decomposing the Oxygen Signal in the Ocean Interior: Beyond Decomposing Organic Matter

“…This decomposition is similar to the recent study by Cliff et al. (2021), although what is referred to as “O 2, eq ” in that study is further decomposed here into . While hydrostatic pressure, subsurface heating, and mixing of water masses with differing temperature and salinity (including diapycnal mixing) affect saturation (Emerson et al., 2012; Enns et al., 1965; Henning et al., 2006; Ito & Deutsch, 2006; Ito et al., 2007; Klots, 1961), physical processes will not impact the total O 2 concentration , which is conservative below the surface for inert gases (Hamme & Emerson, 2002).…”

“…The subscript " E rem" refers to the remineralization. This decomposition is similar to the recent study by Cliff et al (2021), although what is referred to as "O 2, eq " in that study is further decomposed here into…”

“…Using a global ocean circulation-biogeochemical model with explicit TOU, Ito et al (2004) showed that the AOU bias associated with surface O 2 disequilibrium at high latitudes can be significant. Cliff et al (2021) , respectively). In their data-constrained model, ∼11% of the reduction in the ocean's inventory of O 2 from its equilibrium value is due to biological disequilibrium, whereas ∼3% is from physical disequilibrium.…”

“…The aforementioned discussion has implications for estimates of remineralization ratios and preformed and regenerated properties, which have been used to quantify the efficiency of the biological pump (e.g., Ito & Follows, 2013). Using the same data-constrained physical-biogeochemical model as Cliff et al (2021), Khatiwala et al (2019) found that the AOU overestimated the inventory of remineralized carbon by almost 50%. One direct consequence of the work presented here is that 2 bio [ ] E O can reduce this bias.…”

The atmospheric bridge communicated the <i>δ</i><sup>13</sup>C decline during the last deglaciation to the global upper ocean

“…The cGENIE Earth system model is based on a 3-D frictional geostrophic ocean circulation component, plus dynamic and thermodynamic sea ice components, and is configured here at a 36 × 36 horizontal grid resolution with 16 vertical layers in the ocean. The configuration we employ here lacks a dynamical (GCM) atmosphere, with atmospheric transport fixed and provided via a 2-D energy-moisture balance model (Edwards and Marsh, 2005). The low-resolution ocean component and highly simplified atmospheric component make cGENIE much less computationally expensive to run than LOVECLIM and facilitate multiple sensitivity experiments run to (deep-ocean circulation) steady state to help partition and attribute carbon sources and pathways.…”