A song of ice and friction: the impact of basal friction and proglacial lakes on Pleistocene glacial cycles

During the Mid Pleistocene Transition (MPT; 1.2-0.7 Ma) the periodicity of glacial cycles changed from 40 ka to ~100 ka, without a coinciding change in orbital forcing. The MPT therefore results from feedback and changes in the climate system and ice dynamics triggered by the changes in radiative forcing. However, it remains unclear which physical processes are critical for the transition.

Here we explore the role of basal sliding and glacial isostatic adjustment (GIA) in the MPT. Basal sliding is thought to have changed across the MPT due to the erosive action of the ice sheets gradually removing the regolith cover and exposing the underlying bedrock, therefore increasing the friction at the base. GIA modulates this effect by enabling the formation of large proglacial lakes, changing the ice margin from a land-based to a marine environment. We simulate the evolution of the Northern Hemisphere ice sheets during the past 1.5 million years, using an ice-sheet model forced by a climate matrix method.

We show that changing the basal friction has an effect on glacial terminations and consequentially glacial cycle periodicity. Larger friction leads to thicker ice sheets that are more likely to survive a climatic optimum. However, we show that using an unchanging friction coefficient through the Pleistocene, our model still produces change from 40 ka to 100 ka periodicities signifying the MPT. This suggest that the regolith hypothesis is not necessary to explain the MPT.

In addition, we show that the formation of proglacial lakes is required for achieving a full deglaciation of the large Late Pleistocene ice sheets. Ice that floats on water experiences no friction at the base, resulting in high ice velocities. This results in more ice in lower regions and enhances the melt of ice. Here, we find a strong modulating role of GIA. When neglecting bedrock adjustment, thus preventing the formation of large proglacial lakes, we fail to simulate a full deglaciation.