The MMCO-EOT conundrum: Same benthic δ18O, different CO2

Knowledge on climate change during the Cenozoic largely stems from benthic δ18O records, which document combined effects of deep-sea temperature and ice volume. Information on CO2 is expanding but remains uncertain and intermittent. Attempts to reconcile δ18O, sea level, and CO2 by studying proxy data suffer from paucity of data and apparent inconsistencies among different records. One outstanding issue is the difference suggested by proxy CO2 data between the Eocene-Oligocene boundary (EOT) and the Middle-Miocene Climatic Optimum (MMCO), while similar levels of δ18O are shown during these times. This conundrum implies changing relations between δ18O, CO2, and temperature over time. Here we use a coupled climate-ice sheet model, forced by two different benthic δ18O records, to obtain continuous and mutually consistent records of δ18O, CO2, temperature, and sea level over the period 38 to 10 Myr ago. We show that the different CO2 levels between the EOT and MMCO can be explained neither by the standard configuration of our model nor by altering the uncertain ablation parametrization on the East Antarctic Ice Sheet. However, we offer an explanation for the MMCO-EOT conundrum by considering erosion and/or tectonic movement of Antarctica, letting the topography evolve over time. A decreasing height of the Antarctic continent leads to higher surface temperatures, reducing the CO2 needed to maintain the same ice volume. This also leads to an increasing contribution of ice volume to the δ18O signal. This result is, however, dependent on how the topographic changes are implemented in our ice sheet model