Multivariate data assimilation in a seamless sea ice prediction system based on AWI-CM

We implement multivariate data assimilation in a seamless sea ice prediction system based on the fully-coupled AWI Climate Model (AWI-CM, v1.1). AWI-CM has an ocean/ice component with unstructured-mesh discretization and smoothly varying spatial resolution, which aims for seamless sea ice prediction across a wide range of space and time scales. The assimilation uses a Local Error Subspace Transform Kalman Filter coded in the Parallel Data Assimilation Framework. To test the robustness of the assimilation system, a perfect-model experiment is configured to assimilate synthetic observations. Real observations from sea ice concentration, thickness, drift, and sea surface temperature are further assimilated in the system. The analysis results are evaluated against independent in-situ observations and reanalysis data. Further experiments that assimilate different combinations of variables are conducted to understand their individual impacts on the analysis step. Particularly we find that assimilating sea ice drift improves the sea ice thickness estimate in the Antarctic, and assimilating sea surface temperature is able to avert a circulation bias of the free-running model in the Arctic Ocean at mid-depth. We also test the performance of an extended experiment where the atmosphere is constrained by nudging toward reanalysis data. The second version of the system assimilating more observations also with a new atmospheric model is currently under development.