Automating Sea Ice Characterisation from X-Band SAR with Co-Located Airborne Laser Scanner Data Obtained During The MOSAiC Expedition

The research vessel ‘Polarstern’, moored to an ice floe, completed a year long drift with Arctic pack ice in the autumn of 2020. During that expedition, named MOSAiC, a comprehensive data set of airborne laser scanner (ALS) and spaceborne X-band SAR images in the area of the research vessel was acquired. With successful fusion of these two measurements, we can extrapolate sea ice features from the ALS data to the entire SAR scene using a convolutional neural network (CNN). From two preliminary scenes of ALS data we are able to show this for classes of sea ice roughness. This will be the basis for more comprehensive research, once the complete data set is available.