Using ground-penetrating radar to determine the representativeness of ice core surface mass balance records at ice rises along the Princess Ragnhild Coast, East Antarctica

Several studies have shown that there is often a poor match between surface mass balance (SMB, mass gain at the surface of the ice sheet) simulated by regional climate models and the one locally measured from ice cores in Antarctica. Models’ representation of the physical processes that affect SMB is known to be imperfect, while ice core records may be strongly influenced by local processes such as post-depositional wind redistribution and precipitation intermittency. These two sources of uncertainty likely both have a role to play in the discrepancy identified between modeled and observed ice core SMB estimates over the past centuries.

The goal here is to estimate the uncertainties associated with the difference between a point-wise measurement of SMB as provided by the ice core and the SMB averages over a grid of several square kilometers of the models. To do so, we use ground-penetrating radar (GPR) data, collected over several ice rises, located along the high accumulation Princess Ragnhild Coast (East Antarctica), to obtain a multi-year resolution record that goes back ∼30-40 years, representing SMB spatial and temporal variability at the scale of a few km² for each ice rise. Ice cores were collected during each radar field campaign, which allows us to place age constraints on the radar stratigraphy obtained and compare the GPR SMB estimates with the ice core SMB estimate.

Therefore, we are able to calculate an error of representativeness for each ice core SMB, estimated as the difference between the average GPR SMB over a few km² and the ice core SMB. This representativeness error can be split into two components: a systematic error (on the order of ∼0.1 m w.e. yr^-1) and a random error (on the order of ±1 cm w.e. yr^-1). Finally, we then compare our corrected ice core SMB records to regional SMB derived from a state-of-the-art polar-oriented regional climate model to quantify the impact of ice core uncertainties on the modeled-observed SMB discrepancy.