Net effect of ice-sheet–atmosphere interactions reduces simulated transient Miocene Antarctic ice-sheet variability

Benthic δ18O levels vary strongly during the warmer-than-modern early and mid-Miocene (23 to 14 Myr ago), suggesting a dynamic Antarctic ice sheet (AIS). So far, however, realistic simulations of the Miocene AIS have been limited to equilibrium states under different CO2 levels and orbital settings. Earlier transient simulations lacked ice-sheet–atmosphere interactions and used a present-day rather than Miocene Antarctic bedrock topography. Here, we quantify the effect of ice-sheet–atmosphere interactions, running the ice-sheet model IMAU-ICE using climate forcing from Miocene simulations by the general circulation model GENESIS. Utilising a recently developed matrix interpolation method enables us to interpolate the climate forcing based on CO2 levels (between 280 and 840 ppm), as well as varying ice-sheet configurations (between no ice and a large East Antarctic Ice Sheet). We furthermore implement recent reconstructions of Miocene Antarctic bedrock topography. We find that the positive albedo–temperature feedback, partly compensated for by a negative feedback between ice volume and precipitation, increases hysteresis in the relation between CO2 and ice volume. Together, these ice-sheet–atmosphere interactions decrease the amplitude of Miocene AIS variability in idealised transient simulations. Forced by quasi-orbital 40 kyr forcing CO2 cycles, the ice volume variability reduces by 21 % when ice-sheet–atmosphere interactions are included compared to when forcing variability is only based on CO2 changes. Thereby, these interactions also diminish the contribution of AIS variability to benthic δ18O fluctuations. Evolving bedrock topography during the early and mid-Miocene also reduces ice volume variability by 10 % under equal 40 kyr cycles of atmosphere and ocean forcing.