The need to better understand long-term climate/ice sheet feedback loops ismotivating efforts to couple ice sheet models into Earth System models whichare capable of long-timescale simulations. In this paper we describe acoupled model that consists of the University of Victoria Earth SystemClimate Model (UVic ESCM) and the Pennsylvania State University Ice model(PSUI). The climate model generates a surface mass balance (SMB) field via asub-gridded surface energy/moisture balance model that resolves narrow icesheet ablation zones. The ice model returns revised elevation, surface albedoand ice area fields, plus coastal fluxes of heat and moisture. An arbitrarynumber of ice sheets can be simulated, each on their own high-resolution gridand each capable of synchronous or asynchronous coupling with the overlyingclimate model. The model is designed to conserve global heat and moisture. Inthe process of improving model performance we developed a procedure toaccount for modelled surface air temperature (SAT) biases within theenergy/moisture balance surface model and improved the UVic ESCM snow surfacescheme through addition of variable albedos and refreezing over the icesheet.
A number of simulations for late Holocene, Last Glacial Maximum (LGM), andEemian climate boundary conditions were carried out to explore thesensitivity of the coupled model and identify model configurations that bestrepresented these climate states. The modelled SAT bias was found to play asignificant role in long-term ice sheet evolution, as was the effect ofrefreezing meltwater and surface albedo. The bias-corrected model was able toreasonably capture important aspects of the Antarctic and Greenland icesheets, including modern SMB and ice distribution. The simulated northernGreenland ice sheet was found to be prone to ice margin retreat at radiativeforcings corresponding closely to those of the Eemian or the present-day.