My research group primarily investigates novel approaches to retrieve information from satellite data for improved estimation of sea ice parameters, such as ice concentration. We have been looking into both physically based approaches, for example using radiative transfer models, and data-driven approaches, such as convolutional neural networks. Our retrievals primarily utilize passive and active microwave data, and are carried out with a vision for use in ice operations, such as data assimilation. The use of synthetic aperture radar (SAR) data allows for retrievals at higher spatial resolutions than are possible with passive microwave data. This could be particularly useful for coupling of the sea ice with the atmosphere. Recently, we have started investigating the physics of the stably, stratified boundary layer. These boundary layers form when air flows over a cold surface, such as sea ice. They are characterised by small-scales and intermittent turbulence and are difficult to capture in relatively coarse resolution numerical weather forecasting models.