Abstract
Abstract
The ocean plays a crucial role in the formation and melting of sea ice, which also affects the performance of sea ice data assimilation. To evaluate how ocean model complexity affects the assimilation performance, the sea ice model is separately coupled with two distinct ocean models. In this study, the satellite-derived sea ice concentration and thickness data in the Arctic are assimilated into a coupled sea ice-ocean model, using the ensemble Kalman filter (EnKF) approach. Utilizing a fully dynamic ocean model effectively mitigates ensemble convergence during the Arctic freezing season compared to a simplified ocean model, thereby improving the performance of sea ice data assimilation. However, when only sea ice concentration data are assimilated, an anomalous spatial pattern in sea ice thickness emerges in the Arctic, regardless of the type of ocean model. This abnormal pattern is effectively resolved through the joint assimilation of satellite-derived sea ice concentration and thickness data, leading to improved consistency in Arctic sea ice state estimation.