Risk of Sea Level Rise from Antarctic Ice Sheet Instability

Abstract:

Large parts of the Antarctic ice sheet lie below sea level and may be vulnerable to Marine Ice Sheet Instability (MISI), a positive feedback in which ice shelf collapse or exposure to warmer water triggers self-sustaining ice loss. But uncertainty quantification with perturbed parameter ensembles - for probabilistic projections of the risk of sea level rise under MISI - is currently precluded for models that can explicitly simulate migration of the grounding line (that divides ice resting on bedrock from floating ice shelves), due to their computational expense.

We present a new approach implemented in the ice sheet model GRISLI that does not require high resolution and therefore allows uncertainty quantification: we parameterise grounding line retreat rate as a function of an effective basal drag coefficient. We vary the parameters of this scheme along with 13 other inputs (basal drag law exponent, instability onset by sector, and bedrock topography) to generate a 3000 member ensemble. We calibrate the ensemble in a Bayesian statistical framework using observations of present day mass loss in the Amundsen Sea, where the grounding line is currently retreating. We obtain calibrated numerical model projections of the probability of sea level rise over the next 200 years in the event of Antarctic ice sheet instability.

We find the projections are sensitive to model inputs, such as the basal drag law exponent. Nevertheless, our assessment is consistent with the lower end of previous estimates, indicating that the highest of "upper end" estimates are unlikely. This result is due to both ice dynamical theory, which constrains the regions over which MISI can occur and the maximum tensile stresses at the grounding line, and calibration, which constrains the maximum grounding line retreat rate and the values of effective basal drag coefficient over which this can occur. Our results highlight the importance of formal quantification of the effect of model uncertainties, and of constraining projections to plausible values using ice dynamical theory and observational calibration.