摘要:This paper evaluates the simulation of Arctic sea ice states using an ocean-ice coupled model that employs LASG/IAP (the State Key Laboratory of Numerical Modeling for Atmospheric Sciences and Geophysical Fluid Dynamics/the Institute of Atmospheric Physics) Climate Ocean Model (LICOM) and the sea-ice model from the Bergen Climate Model (BCM). It is shown that the coupled model can reasonably reproduce the major characteristics of the mean state, annual cycle, and inter-annual variability of the Arctic sea ice concentration. The coupled model also shows biases that were generally presented in other models, such as the underestimation of summer sea ice concentration and thickness as well as the unsatisfactory sea ice velocity. Sensitivity experiments indicate that the insufficient performance of the ocean model at high latitudes may be the main reason for the biases in the coupled model. The smoother and the fake “island”, which had to be used due to the model's grid in the North Pole region, likely caused the ocean model's weak performance. Sea ice model thermodynamics are also responsible for the sea ice simulation biases. Therefore, both the thermodynamic module of the sea ice component and the model grid of the ocean component need to be further improved.