Greenland Ice Sheet evolution during the Last Interglacial with an improved surface mass balance modeling approach

The Last Interglacial period (LIG) (130 - 116 kaBP), characterized by higher global mean temperature and sea levels compared to the present-day due to the Earths orbit configuration, has been well-studied as a recent example of a climate period warmer than today. There is particular interest in studying the ice sheet-climate interactions in view of our current climate change. However, the extent of the ice sheet and its contribution to the rise of sea levels during the LIG remain debatable as different approaches suggest a wide range of estimations. In order to cover such a long period, some processes are simplified in the modeling approach by using prescribed forcings, simple surface mass balance (SMB) schemes, or equilibrium simulations, which all affect the numerical estimation of ice sheet evolution. In our work, to perform transient simulations, we use an Earth system model of intermediate complexity (iLOVECLIM), which has been widely used to study various long-timescale periods. Additionally, we use a physically-based energy and mass balance model with 15 vertical snow layers BESSI (BErgen Snow Simulator) to account for the effect of insolation changes as well as snow-albedo feedback. The climate forcings of the snow model are obtained by running iLOVECLIM transiently from 135 to 115 kaBP, downscaled over the Northern Polar region. Using the SMB computed by BESSI, we then simulate the ice sheet evolution during the LIG with GRISLI - the ice sheet model in the iLOVECLIM framework. To assess the benefits of using a physically-based SMB model in the ice sheets simulation, the outputs of GRISLI-BESSI are compared to the current SMB scheme of iLOVECLIM, a simple parametrization called ITM (Insolation Temperature Melt). The Greenland ice sheet volume simulated by the two SMB models reaches the minimum value at 127.5 kaBP, around 500 years after the peak of global mean temperature. The magnitude of ice sheet retreat and its contribution to the sea level in ITM simulations are significantly higher than in BESSI due to an overestimation of the zones of ablation. The findings suggest that, compared to a parameterization, we have more confidence in the ice sheet estimation with a physically-based SMB model. Further works with fully interactive ice sheet modeling that takes into account the melt-elevation feedback can improve the simulation of the ice sheet-climate interactions of long-time scales.