A smaller glacial ocean carbon inventory?

<p>Previous studies attempting to explain Pleistocene atmospheric CO<sub>2</sub> variations have focused on mechanisms that transfer carbon (C) between the oceanic, atmospheric and terrestrial reservoirs, with the underlying assumption that the total C inventory in these three Earth’s surface reservoirs remained constant during glacial-interglacial cycles. Under this framework, ocean C inventory would have been marginally increased by 500-1000 GtC (1-2%) during the glacial period. Here, we show that past ocean C inventory can be revealed by reconstructed bulk ocean <sup>14</sup>C/<sup>12</sup>C (denoted as ∆<sup>14</sup>C) and atmospheric <sup>14</sup>C production rates with an Earth system model - cGENIE. First, we develop a bulk ocean ∆<sup>14</sup>C record that spans the last 40 ka from thousands of benthic foraminifera and deep sea coral ∆<sup>14</sup>C data with a fairly good coverage of the global seafloor. We then run cGENIE under constant pre-industrial boundary conditions, with the only forcing being atmospheric <sup>14</sup>C production rates reconstructed by geomagnetic field intensity records and ice core record of <sup>10</sup>Be fluxes. Under most of the <sup>14</sup>C production scenarios, the simulated bulk ocean ∆<sup>14</sup>C are significantly lower than our composite during the Last Glacial Maximum as well as the early deglaciation. Bulk ocean ∆<sup>14</sup>C is a metric controlled by <sup>14</sup>C production rates and ocean C inventory, with the state of ocean circulation playing a minor role.  Our finding suggests either glacial <sup>14</sup>C production was much higher and/or glacial C inventory was much lower than previously thought. Implications of both possibilities are discussed. In particular, the second possibility highlight the exchange of C and ALK between Earth’s surface and geological reservoirs as a critical missing piece in searching for a complete theory of glacial-interglacial atmospheric CO<sub>2 </sub>variability.</p>