Approach

“Lakes4Antarctica” will use Earth Observation (EO) data provided by ESA’s Copernicus programme allowing for an accurate and spatio-temporally detailed monitoring of the global ice sheets, as performed as part of several ESA projects including the ESA Ice sheet Mass Balance Inter-comparison Exercise (IMBIE), the ESA CCI Antarctic Ice Sheets (AIS) project, the ESA CCI Sea Level Budget Closure (SLBC) project or the 4DAntarctica project. “Lakes4Antarctica” will mainly exploit Sentinel-1 Synthetic Aperture Radar (SAR) and optical Sentinel-2 data in order to address the pressing need for a large-scale and temporally detailed monitoring of Antarctic supraglacial lake extents. We will advance already existing methods for Antarctic supraglacial lake extent derivation developed at DLR using state-of-the-art machine learning and deep learning. Particular focus during algorithm development will be the spatio-temporal transferability of methods in order to enable the computation of Antarctic-wide mapping products in the long-term. As part of “Lakes4Antarctica”, mappings will be generated at intra-annual temporal resolution, i.e. with unprecedented bi-weekly and monthly temporal coverage. Specific focus will be on regions with frequent lake recurrence including large coastal sections along the Antarctic Peninsula with George VI, Larsen C, Wilkins and Bach ice shelves. Further, ice shelves in the Queen Maud, Enderby, MacRobertson and Wilhelm II regions in East Antarctica will be investigated. A thorough performance evaluation of established methods will be undertaken and is key for reliable and accurate mappings. The derived and evaluated mapping products will then be used for integration into modelling, e.g. of hydrological routing and lateral meltwater transport across Antarctic ice shelves, as well as for improvement of the representation of these processes in the RACMO (Regional Atmospheric Climate MOdel) surface mass balance model in the long-term.