Simulating the Antarctic and Greenland ice sheets on millennial timescales using a coupled ice sheet/climate model: sensitivity to climate model bias, refreezing and albedo in Eemian, Last Glacial Maximum, and late Holocene climates

The need to resolve large-scale climate-ice sheet feedbacks is motivating efforts to couple full ice sheet models to Earth System models that are capable of ensembles of multi-millennial simulations of climate change. To this end we have coupled the University of Victoria Earth System Climate Model (UVic ESCM) to the Pennsylvania State Ice Sheet Model (PSUI). The UVic ESCM is an ‘Earth model of intermediate complexity’ (EMIC) that has been extensively used for centennial-to-millennial studies of the Earth System and is under current development and use. PSUI is an ice sheet model that resolves grounded ice, floating ice shelves and a dynamic grounding line. The two models are coupled across the ice-atmosphere interface using an energy moisture balance model on subgrid elevation bins that can correct for model surface air temperature bias. Precipitation is delivered to the ice sheets by the overlying vertically integrated advective/diffusive atmosphere model. Surface evaporation, runoff, and associated heat fluxes are routed to the atmosphere and ocean, respectively. Ice shelves shade the underlying ocean from heat, momentum and moisture fluxes. The ocean receives heat and moisture fluxes from prescribed basal shelf melting and calving. The overall system is designed to conserve heat and moisture to machine precision. Multiple independent ice sheet model grids within a global climate simulation are made possible through use of Fortran derived data types and pointers. The model can run in `synchronous’ or `synchronous accelerated’ modes depending on the length and character of the simulation. The current coupled model is set up to simulate the Antarctic and Greenland Ice Sheets (AIS/GIS) at 20 km resolution within a global climate simulation. A suite of equilibrium simulations has been carried out for representative Eemian, Last Glacial Maximum (LGM) and late Holocene climates, in order to assess the effect of climate model bias and refreezing and albedo schemes on long-term ice sheet evolution. Over millennial timescales, each of these factors can significantly influence ice sheet geometry and ice sheet response to climate perturbations. Long-term ice sheet sensitivity to the representation of modelled surface processes increases as ablation zones and melt extent areas grow, indicating that multi-millennial predictive simulations of the Greenland and Antarctic ice sheets require accurate descriptions of surface processes in addition to proper resolution of ice dynamics and basal processes.