Impact of marine productivity on atmospheric pCO2 during the Last Glacial Maximum: a model-data comparison

Measurements of the air trapped in Antarctic ice cores reveal that atmospheric CO₂ concentration (pCO₂) during the Last Glacial Maximum (LGM) was about 80 ppmv lower than that recorded during the current Holocene interglacial (Bereiter et al., 2015). Studies also show a strong link between pCO₂, ice volume and Antarctic temperature, suggesting pCO₂ as a forcing or amplifying factor behind glacial/interglacial cycles (Petit et al., 1999; Parrenin et al., 2013). Despite such importance in the global climate changes, mechanisms behind rapid variations in pre-anthropic pCO₂ remain elusive. Numerical models emphasized the crucial role of exported marine productivity P_exp, (namely, the Soft Tissue Pump) in such changes. In particular, they feature marine productivity patterns from the Southern Ocean and show that a decrease in P_exp in the Sub-Antarctic zone, linked to a reduction in iron inputs from aeolian dusts, could have increased pCO₂ by 20 to 50 ppmv (Köhler and Fischer, 2006; Martínez-Garcia et al., 2009; Lambert et al., 2012). However, these studies are usually compared to proxy data from the Atlantic sector of the Subantarctic Zone i.e., an area under the direct influence of wind fields that makes it possible to test the “Fe-hypothesis” (Martin et al., 1990) but that is not necessarily representative of the entire ocean (e.g. Lambert et al., 2015). Due to a lack of recent P_exp data compilation but also of direct comparisons with model outputs integrating marine biogeochemistry­­, it remains difficult to understand the role marine biological productivity exerted on the carbon cycle and more specifically on the low pCO₂ during the LGM.

The aim of this study is to explore P_exp patterns during the LGM compared to the pre-industrial Holocene and understand the mechanisms driving their global changes, in order to try and estimate the contribution of marine productivity to the pCO₂ signalbased on (i) a new compilation of P_exp proxy data using the strategy previously proposed by Kohfeld et al. (2005) after Bopp et al., (2003), and (ii) a comparison of these data to outputs from the iLOVECLIM intermediate complexity.

Proxy data show that P_exp is generally higher during the LGM compared to the pre-industrial Holocene. This is particularly the case in the sub-Antarctic and sub-Arctic areas, in the equatorial Atlantic Ocean and in coastal upwelling settings i.e., regions that usually witness higher nutrient content due to revigorated ocean circulation and/or intensified surface winds. Simulations generally confirm such features except from the coastal upwelling and the Southern Ocean, due to a lack of spatial resolution and of aeolian inputs in the model, respectively. However, preliminary results from sensitivity tests show (i) net marine productivity fronts around ~40°N and 45°S due to extended sea ice cover and reduced global temperature, (ii) a decreased P_exp in the Pacific Ocean due to an overall thermohaline circulation slow down and (iii) an increase of P_exp in areas where fertilization by iron-rich dusts is expected (Lambert et al., 2021).