Creating daily high-resolution maps of surface meltwater in the (Ant-)Arctic with deep learning-based superresolution models.- Working with geospatial data, such as, passive microwave, MODIS, digital elevation maps and synthetic aperture radar.- Adapting and benchmarking various deep learning-based image segmentation models, based on UNet, for our task of superresolution on geospatial data.- Optimizing the machine learning models on CPU and GPU clusters on supercomputers- Creating different baseline models.