The calving constraints on inception of Quaternary ice sheets

Abstract

Most numerical models that simulate the inception of Quaternary ice sheets nucleate the ice sheets in high-latitude highlands of the Northern Hemisphere. These nuclei are initially plateau or island ice caps that then expand and merge with one another to become an ice sheet. Before merger, however, these ice caps often must cross continental shelves, seas, or embayments (e.g., Barents Sea, Baltic Sea, Foxe Basin, Hudson Bay) where calving along the ice margins could prevent merger. Therefore, the models must either ignore calving and treat the ice caps as if they were advancing across dry land or include a calving parameter that can be adjusted to allow a grounded or floating ice front to advance across a water body despite calving. The first solution to this problem is unrealistic because when the ice caps reach shorelines their mass is insufficient to have lowered sea level enough to make the shorelines retreat at the rate of advancing ice. The second solution is also unrealistic because, as ice advances into water, slab calving can prevent an ice shelf from forming and, even if it does form, combined slab and tabular calving along its advancing front can prevent the ice shelf from thickening enough to become grounded in the body of water. Calving mechanisms for tabular icebergs and ice slabs are presented to demonstrate that rates of ice calving alone are high enough to prevent ice caps from crossing the water bodies where Quaternary ice sheets had become grounded. Calving rates produced by the models are compatible with the observed calving rates from Alaskan tidewater glaciers and Greenland outlet glaciers along both grounded and floating calving fronts. In this case, the only way by which highland or island ice caps could cross these water bodies would be if sea ice on the water bodies had thickened and grounded independently from the advancing ice caps. This has been called the Marine Ice Transgression Hypothesis.