C51C-0733
Where is the evidence of past collapse of the West Antarctic Ice Sheet?

Friday, 18 December 2015
Poster Hall (Moscone South)
Gail Gutowski, University of Texas at Austin, Austin, TX, United States
Abstract:
Sea level rise estimates from the Last Interglacial period suggest collapse of part of the Antarctic ice sheet. However, there is no direct evidence of this from the ice sheet itself. Englacial layers in ice sheets, sampled directly by ice core drilling and indirectly by ice-penetrating radar, reveal a significant amount about glacial change over time and may contain a signature of the last ice sheet collapse.

We hypothesize there is evidence of ice sheet instability where the observed englacial record deviates from that expected for a steady state WAIS simulated using ice sheet models. However, discrepancies between modeled steady state and observed englacial layer geometry are confounded by uncertainties in model boundary conditions, observed layer ages, and model parameters. To know where the signal of collapse may be best preserved, we must account for the affect of these uncertainties on layer geometry.

We present several tests quantifying the sensitivity of simulated layer geometry to changes in model boundary conditions. We look to areas where englacial geometry has low sensitivity to uncertain boundary conditions to provide the largest signal of ice sheet instability. Where simulated layer geometry is responding strongly to uncertain boundary conditions, we are unlikely to be able to discern a signal of past deglaciation. In the latter case, uncertainty in layer geometry may overwhelm the signal of past ice sheet collapse.

We perform the simulations using the Variational Glacier Simulator (VarGlaS), an ice sheet model with the capacity to model the age of englacial isochrones (Figure 1). We use the latest boundary conditions for geothermal flux, basal topography, and surface mass balance to simulate the steady state behavior of englacial layers in the Thwaites Glacier catchment and the Marie Byrd Land dome. Ensembles of model runs sample the uncertainty in each of the boundary conditions, creating a distribution of simulated englacial layers which accounts for boundary condition uncertainty.