Sea ice plays a major role in the global climate as it represents the interface between the ocean and the atmosphere and thus is of great importance for the energy budget of the whole planet. Climate change has caused a significant rise in air temperature during the last decades that has led to a rapid sea ice decline. Until now, sea ice retreat and thinning are underestimated and poorly represented by climate models as many processes are not well understood yet. Further observations of sea ice thickness and extent are required in order to understand the key processes that lead to sea ice transformations in both space and time. In this thesis the sea ice thickness and underside topography of three different areas northeast of the Svalbard archipelago are investigated during the freeze-up period of September and October 2016. In this pilot project, sea ice draft measurements are conducted using an upward-looking multibeam sonar mounted on a remotely operated vehicle for under ice surveys. The data collected are processed using the hydrographic processing system “CARIS Hips”. A new processing workflow has been developed to measure sea ice draft from underneath the ice. It allows the analysis of the data collected by the upward-looking sonar and the pressure sensor, together with many other sensors mounted on the underwater vehicle, in order to directly compute sea ice draft. Sea ice thickness can be calculated from draft measurements assuming isostatic equilibrium. Three-dimensional topographic images of the underside of the sea ice are produced and correlated with the respective sea ice thickness maps. The spatial and vertical resolution of the multibeam sonar is also calculated. Moreover, multibeam sonar derived sea ice thickness datasets are compared to sea ice thickness data collected by an electromagnetic induction sounding device during the same surveys. Finally, the “Freezing-degree days” model is used to assess sea ice thermodynamic growth of the data collected during the field campaign. Snow cover is taken into account in the model thanks to snow depth measurements conducted on the areas with a Magna Probe. It is found that the two instruments for sea ice thickness measurements are in good agreement and have the same vertical resolution. However, the multibeam sonar is found to have a better lateral resolution and to be more accurate than the electromagnetic device when measuring sea ice ridges. The assessment of sea ice thermodynamic growth is hindered by the high spatial variability of the three areas of this campaign. Nonetheless the model predictions are found to be consistent with the formation from open water of a few centimeters of new ice during a survey period of four weeks. This thesis also suggests some improvements to the navigation of the underwater vehicle for sea ice draft measurement purposes and to the multibeam sonar renavigation script. The results of this thesis prove that the new processing workflow implemented in CARIS Hips allows for a reliable, efficient, and high resolution retrieval of sea ice draft measurements collected by an upward-looking sonar mounted on a remotely operated vehicle. The methods presented in this thesis can be adopted for a future year-round spatial and temporal study of sea ice thickness and underside morphology, necessary to fill the existing data gap during winter time in the Arctic. The use of the multibeam sonar together with the many interdisciplinary sensors mounted on the remotely operated vehicle empowers a complete overview of the sea ice underside environment and contributes to the improvement of climate models.