Transient Variability of the Miocene Antarctic Ice Sheet Smaller Than Equilibrium Differences

Abstract

During the early to mid‐Miocene, benthic δ18O records indicate large ice volume fluctuations of the Antarctic ice sheet (AIS) on multiple timescales. Hitherto, research has mainly focused on how CO2 and insolation changes control an equilibrated AIS. However, transient AIS dynamics remain largely unexplored. Here, we study Miocene AIS variability, using an ice sheet‐shelf model forced by climate model output with various CO2 levels and orbital conditions. Besides equilibrium simulations, we conduct transient experiments, gradually changing the forcing climate state over time. We show that transient AIS variability is substantially smaller than equilibrium differences. This reduces the contribution of the AIS to δ18O fluctuations by more than two thirds on a 40‐kyr timescale, hence requiring a larger contribution by deep‐sea‐temperature variability. The growth rates are much slower than the decay rates, which ensures variability around a preferred small state. Finally, if the bedrock topography enlarges the West Antarctic land surface, AIS self‐sustenance increases.

Anti‐Phased Miocene Ice Volume and CO2 Changes by Transient Antarctic Ice Sheet Variability

“…The index is linearly changed with time from 1 to 0 to 1, with periods of 100, 400, and 40 kyr and the conceptual model integration time is adapted to the forcing period and set to 100, 400, and 200 kyr. The growth and decay rates are then calibrated to find a reasonably close correspondence between the transient conceptual model simulations and the PISM simulations of Stap et al (2019): …”

“…The blue shaded area marks where the ice sheet is smaller than equilibrium and the red shaded area where it is larger. (c) Ice thickness from steady‐state simulations of the Miocene Antarctic ice sheet using PISM (Stap et al, 2019), started with a present‐day topography (Bedmap2) after deglaciation and isostatic rebound, for the lowest CO 2 level and (d) for the highest CO 2 level. Intermediate level results are not shown here.…”

“…To apply our conceptual model to the Miocene AIS, we prescribe a C ‐ V eq (in this case CO 2 ‐index‐ V eq ) relation based on ice dynamical model results we presented in an earlier study (Stap et al, 2019) (Figures 2c and 2d). These were obtained by running the 3D thermodynamical PISM version 0.7.3 (Bueler & Brown, 2009; Winkelmann et al, 2011) into equilibrium using Miocene forcing conditions (precipitation, near‐surface atmospheric, and 3D ocean temperatures).…”

“…Using time steps of 100 years, we repeat the transient simulations of Stap et al (2019) with the conceptual model using the same CO 2 ‐index to interpolate between the forcing climates. The index is linearly changed with time from 1 to 0 to 1, with periods of 100, 400, and 40 kyr and the conceptual model integration time is adapted to the forcing period and set to 100, 400, and 200 kyr.…”

“…On the basis of equilibrium simulations, AIS modeling studies have demonstrated the importance of including Antarctic bedrock topography changes (Colleoni et al, 2018) and ice sheet‐climate feedbacks, ice calving parameterizations, and changes in the ice sheet's oxygen isotopic composition (Gasson et al, 2016), for simulating strong ice sheet‐induced benthic δ 18 O variations. However, we found that the slow response time of the AIS to orbital timescale forcing changes significantly reduces the magnitude of ice volume variability and hence its contribution to benthic δ 18 O fluctuations, in transient simulations (Stap et al, 2019).…”