Spatio-Temporal Dynamics of Snow on Central Arctic Sea Ice and its Implications

This thesis presents a comprehensive investigation of the interaction between precipitation and wind-induced erosion and deposition of snow on Arctic sea ice. The study uses observations from the 2019-2020 MOSAiC expedition and makes use of the 3D snow cover- and snow transport model ALPINE3D. Additionally, a series of related studies with my participation is presented. We describe the application of a snow depth-SWE (Snow-Water-Equivalent) function using SnowMicroPen force signals and its application to Magnaprobe snow depth data for MOSAiC transect lines. The model accurately reconstructed SWE compared to direct measurements, allowing for the estimation of the accumulated snow mass between October 2019 and May 2020. The study found that the net accumulation mass in spring was significantly lower than that observed during the earlier SHEBA or N-ICE2015 expeditions, likely due to climate change. The study also found that the snow accumulation rate on first year ice (FYI) was higher than that on second year ice (SYI), and the mass on both ice types approached a saturation point towards the end of February, although SYI had twice the snow mass at the beginning of the winter. In addition, we investigated the plausibility of precipitation sensors measurements installed during MOSAiC, finding that only a small fraction of the data was usable without post-processing. Snowfall sensors installed close to the ice at low altitude tended to overestimate the snowfall rate due to drifting snow falsely detected as precipitation, while sensors on the ship installed at over 20 m height showed almost no such behavior. The study calculated a plausible range of 72 to 107 mm for the total cumulated precipitation mass between October 2019 and May 2020. These values were compared with the accumulated snow amount on the ground to estimate the relative amount of snow that had disappeared after snowfall. We conclude that at least about 50 % of precipitated snow have been eroded and transported away, which is roughly in line with previous studies. Furthermore, the 3D snow cover and snow transport model ALPINE3 was adjusted, calibrated, and applied to simulate snow processes over Arctic sea ice. The model underwent several statistical and qualitative comparisons. ALPINE3D captured the statistical distribution of snow height differences well, but both the statistical and qualitative evaluations showed a lack of elongated snow bed forms such as dunes, which is likely due to the absence of a dynamic mesh. Regarding surface density, the model was compared to measurements, which showed excellent density reproduction up to a certain point but with a slight underestimation at later stages. Additionally, the model showed strong temporal variations in surface density, which were significantly reduced by increasing the fluid threshold at the surface for wind-induced snow transport. Furthermore, we used the model to investigate the deposition and erosion of snow in the area of pressu