Surface Mass Balance Models Vs. Stake Observations: A Comparison in the Lake Vostok Region, Central East Antarctica

Abstract

The surface mass balance (SMB) is very low over the vast East Antarctic Plateau, for example in the Vostok region, where the mean SMB is on the order of 20–35 kg m-2 a-1. The observation and modeling of spatio-temporal SMB variations are equally challenging in this environment. Stake measurements carried out in the Vostok region provide SMB observations over half a century (1970–2019). This unique data set is compared with SMB estimations of the regional climate models RACMO2.3p2 (RACMO) and MAR3.11 (MAR). We focus on the SMB variations over time scales from months to decades. The comparison requires a rigorous assessment of the uncertainty in the stake observations and the spatial scale dependence of the temporal SMB variations. Our results show that RACMO estimates of annual and multi-year SMB agree well with the observations. The regression slope between modelled and observed temporal variations is close to 1.0 for this model. SMB simulations by MAR are affected by a positive bias which amounts to 6 kg m-2 a-1 at Vostok station and 2 kg m-2 a-1 along two stake profiles between Lake Vostok and Ridge B. None of the models is capable to reproduce the seasonal distributions of SMB and precipitation. Model SMB estimates are used in assessing the ice-mass balance and sea-level contribution of the Antarctic Ice Sheet by the input-output method. Our results provide insights into the uncertainty contribution of the SMB models to such assessments.