Coupled climate-carbon cycle simulation of the Last Glacial Maximum atmospheric CO<sub>2</sub> decrease using a large ensemble of modern plausible parameter sets

Abstract

Abstract. During the Last Glacial Maximum (LGM), atmospheric CO 2 was around 90 ppmv lower than during the preindustrial period. Despite years of research, however, the exact mechanisms leading to the glacial atmospheric CO 2 drop 15 are still not entirely understood. Here, a large (471-member) ensemble of GENIE-1 simulations is used to simulate the equilibrium LGM minus preindustrial atmospheric CO 2 concentration difference (∆CO 2 ). The ensemble has previously been weakly constrained with modern observations and was designed to allow for a wide range of large-scale feedback response strengths. Out of the 471 simulations, 315 complete without evidence of numerical instability, and with a ∆CO 2 that centres around -20 ppmv. Roughly a quarter of the 315 runs predict a more significant atmospheric CO 2 drop, between ~30 and 90 20 ppmv. This range captures the error in the model's process representations and the impact of processes which may be important for ∆CO 2 but are not included in the model. These runs jointly constitute what we refer to as the "plausible glacial atmospheric CO 2 change-filtered (PGACF) ensemble".Our analyses suggest that decreasing LGM atmospheric CO 2 tends to be associated with decreasing SSTs, increasing sea ice 25 area, a weakening of the Atlantic Meridional Overturning Circulation (AMOC), a strengthening of the Antarctic Bottom Water (AABW) cell in the Atlantic Ocean, a decreasing ocean biological productivity, an increasing CaCO 3 weathering flux, an increasing terrestrial biosphere carbon inventory and an increasing deep-sea CaCO 3 burial flux. The increases in terrestrial biosphere carbon are predominantly due to our choice to preserve rather than destroy carbon in ice sheet areas. However, the ensemble soil respiration also tends to decrease significantly more than net photosynthesis, resulting in relatively large 30 increases in non-burial carbon. In a majority of simulations, the terrestrial biosphere carbon increases are also accompanied by decreases in ocean carbon and increases in lithospheric carbon. In total, however, we find there are 5 different ways of 2Comparison of the PGACF ensemble results against observations suggests that the simulated LGM physical climate and biogeochemical changes are mostly of the right sign and magnitude or within the range of observational error, except for the change in global deep-sea CaCO 3 burial flux -which tends to be overestimated. We note that changing CaCO 3 weathering flux is a variable parameter (included to account for variation in both the CaCO 3 weathering rate and the un-modelled CaCO 3 shallow water deposition flux), and this parameter is strongly associated with changes in global CaCO 3 burial rate. The 5 increasing terrestrial carbon inventory is also likely to have contributed to the LGM increase in deep-sea CaCO 3 burial flux via the process of carbonate compensation. However, we do not yet rule out either of these processes as causes of ∆CO 2 since missing processes such as Si fertilisation, Si leakage and the effect of d...

Low terrestrial carbon storage at the Last Glacial Maximum: constraints from multi-proxy data

“…Vecsei and Berger (2004) reconstructed coral reef growth history and provided a lower-limit estimate of about 380 GtC for the amount of shallow water carbonate deposition over the deglacial period. Other estimates of CaCO 3 deposition amount to 1200 GtC or even more (Milliman, 1993;Kleypas, 1997;Ridgwell et al, 2003). This yields a large range of possible scenarios with substantial impacts on atmospheric CO 2 .…”

“…The transfer of 13 C and carbon between the oceanatmosphere system and reactive sediments and the lithosphere is affected by the well-documented reorganization of the marine carbon cycle across the deglaciation (e.g., Sigman and Boyle, 2000;Sigman et al, 2010;Fischer et al, 2010;Elderfield et al, 2012;Ciais et al, 2013;Martinez-Garcia et al, 2014;Jaccard et al, 2016;Cartapanis et al, 2016). The reorganization includes changes in the following: CO 2 solubility, ocean ventilation (e.g., Burke and Robinson, 2012;Sarnthein et al, 2013;Skinner et al, 2017), and water mass distribution (e.g., Duplessy et al, 1988;Peterson et al, 2014;Lippold et al, 2016;Menviel et al, 2017;Gottschalk et al, 2018); air-sea gas transfer rates, for example via changes in sea-ice extent (e.g., Stephens and Keeling, 2000;Gersonde et al, 2005;Waelbroeck et al, 2009;Sun and Matsumoto, 2010), export (e.g., Kohfeld et al, 2005;Jaccard et al, 2013), and the remineralization (e.g., Bendtsen et al, 2002;Matsumoto, 2007;Taucher et al, 2014) of biogenic material associated with the cycling of organic carbon, CaCO 3 , and biogenic opal (Tschumi et al, 2011;Roth et al, 2014;Matsumoto et al, 2014); coral reef growth (e.g., Berger, 1982;Milliman and Droxler, 1996;Kleypas, 1997;Ridgwell et al, 2003;Vecsei and Berger, 2004; and the input of dust, nutrients, and lithogenic material by atmospheric deposition (e.g., …”

A large increase in the carbon inventory of the land biosphere since the Last Glacial Maximum: constraints from multi-proxy data

“…8A-D for changes in organic weathering and land carbon). Indeed, changes in these reservoirs have been invoked as an alternative explanation for the reconstructed deglacial change in marine δ 13 C (Broecker, 1982;Zimov et al, 2006Zimov et al, , 2009Zech et al, 2011;Kemppinen et al, 2018). Here, we represent variations in these carbon stocks by varying weathering-burial imbalances of organic material in addition to adding variable amounts of isotopically light land carbon to the atmosphere in our 500,000 member ensemble.…”

“…a larger terrestrial carbon inventory during the glacial (Zimov et al, 2006(Zimov et al, , 2009Zech et al, 2011;Kemppinen et al, 2018), to 1,500 GtC (Adams and Faure, 1998) associated with its own uncertainties and limitations. Studies suggesting a negative ∆land (Zimov et al, 2006(Zimov et al, , 2009Zech et al, 2011;Kemppinen et al, 2018) argue that the reconstructed deglacial δ 13 C DIC change is due to changes in other organic carbon reservoirs. These postulated changes would not only need to explain the observed deglacial change in δ 13 C DIC , but in addition also explain the imprint of a negative ∆land on the isotopic budget.…”