GPR for rapid and accurate snow-thickness measurements on Antarctic sea ice

Snow thickness on sea ice is not known on the regional scales and to the accuracy required for climate analysis and to assess its significant role in the Earth's cryosphere mass balance. The thickness of snow on sea ice is a largely under-sampled, but crucially important parameter when estimating sea-ice thickness from space-borne sensors. Traditional sampling methods comprise meter-stick measurements or snow probes, which yield highly accurate data but are sparse in space and time, whereas remote sensing data have the potential to cover large areas but exhibit large uncertainties. To increase the local data volume and to apply as an intermediate level for remote-sensing data validation, we present non-destructive ground-penetrating radar (GPR) snow-thickness data of remarkable accuracy (millimetre to centimetre) when compared to in-situ data. The semi-regional data collected during two Antarctic campaigns in late winter/early spring confirm the underestimation of snow thickness on sea ice using remote-sensing methods by roughly 30%. The radar wave propagation was found to be rather constant in snow (+/- 6%), supporting the applicability of GPR for regional snow thickness surveys of snow thicker than 10cm, which was the detection limit of the off-the-shelf GPR setup used in this study.