Antarctic firn aquifers detection algorithm based on satellite and regional climate model data

<p>In the past decade, localized in-situ observation and radar measurements have revealed the presence of firn aquifers on the Antarctic Ice Sheet. They are believed to be an important component of the hydrological system of the ice sheet, but currently no large-scale observational studies exist, hampering a detailed assessment of their contribution to runoff and sea-level rise.<br />We present a probability map of firn-aquifer presence around the Antarctic Ice Sheet, based on a combination of remote sensing and climate modeling data. C-band radar imagery from the Sentinel-1 and Advanced Scatterometer (ASCAT) missions, together with climate data from the regional atmospheric climate model RACMO2.3p2, are combined to map the probability of detecting seasonal and perennial firn aquifers at 5.5x5.5 km² resolution in the period 2017 to 2020.<br />Our method is based on Monte Carlo simulations: its algorithm predicts the probability of aquifer presence based on a set of fixed inputs, to which dedicated thresholds and weights are assigned.<br />In agreement with observation from previous studies, we find a high probability of firn aquifer presence in the north and northwest of the Antarctic Peninsula, and on the Wilkins and George VI ice shelves. A low probability is found in the higher elevation areas of the central Antarctic Peninsula, where the presence of aquifers is not expected due to the absence of surface melt. Thanks to the large spatio-temporal availability of the input datasets, the presence of aquifers in different regions of Antarctica has been estimated. The methodology has been validated in selected regions of Greenland, where the presence of aquifers has been observed using in-situ and remotely sensed data.</p>