Deep learning-based supraglacial lake extent and depth detection on the Greenland Ice Sheet by combining ICESat-2 and Sentinel-2 data

Abstract:
Supraglacial lakes are an important parameter for understanding the current and future state of the Greenland Ice Sheet. Previous studies have focused on mapping supraglacial lake extent using optical and radar imagery, while lake depth is more difficult to estimate due to sparse temporal and spatial coverage of laser altimeters such as ICESat-2. We present a supervised deep learning approach to predict lake extent and depth based on the subtle spectral signatures acquired from Sentinel2 imagery. The model is trained on an existing lake extent product and elevation profiles derived from ICESat-2. The output of this approach is a proof-of-concept study whereby deep learning can utilise contextual information from the input image to produce a lake depth and extent prediction. The preliminary results show that the methodology is feasible as the output model successfully produces a reasonable lake extent and depth prediction despite data limitations. This work forms part of the European Space Agency’s Greenland Ice Sheet Climate Change Initiative (ESA GIS CCI+), which runs from December 2022 until 2025.