【摘要】 利用宇如聪等１９９５年建立的北极区域冰洋耦合模式，以１９６６～１９９１年期间逐月的月平均实测海平面气温和气压场为强迫场，模拟了上述２６年间北极海冰的时间演变和空间分布，着重分析了大西洋及欧洲沿岸一侧的巴伦支海和格陵兰海的海冰状况，并与目前能够得到的北极海冰密集度观测资料做了对比，结果表明：（１）模式对巴伦支海海冰年际变化的模拟是比较成功的，表现在不仅模拟的１９６９～１９７９和１９７９～１９８７这两个时段的主要变化趋势和观测事实比较一致，而且模拟出了１９７９和１９８４这两个多冰和少冰的极端年份。模拟的主要年际变化出现在巴伦支海东部和中部海域，这一点同观测事实是一致的，不过模拟的年际变化偏于新地岛西侧，而观测结果则更靠近挪威沿岸。（２）模式未能在格陵兰海模拟出与观测一致的年际变化。根据巴伦支海和格陵兰海模拟结果的对比有理由推测：巴伦支海的海冰可能更多地受到热力学过程的控制，而动力学因子对格陵兰海海冰的作用则不可忽视。（３）模拟和观测的巴伦支海和格陵兰海海冰的季节循环均滞后于气温的季节循环，但模拟结果滞后的时间更长，事实上模拟的冬季海冰极值比观测滞后１～２个月，而夏季无冰期比观测结果长两个月以上，这是模式需要改?更多还原

【Abstract】 The variations of sea ice in the Arctic Ocean during 1966 to 1991 are simulated by using an Arctic Ocean circulation and thermodynamic sea ice model developed by Yu Rucong et al. in 1995. The model is run with boundary conditions from observed monthly sea surface air temperature and pressure. Analyses are focused on the simulation of sea ice in the Barents and Greenland seas where the fluctuations of sea ice are more significant than in other regions. The results indicate: (1) The simulation of sea ice in the Barents Sea is successful, reflecting not only the interannual variability during 1969 to 1987, but also two extreme events (i.e. sea ice is extremely heavy in 1979 and extremely light in 1984) which are in good agreement with observation. The simulated major annual variability appeared in the east and center of the Barents Sea, which is consistent with observation although the location is somewhat drifted to the west of Newfoundland rather than near the coastline of Norway; (2) The simulation of annual variability of sea ice in the Greenland Sea is not consistent with observation. Therefore it is possible that sea ice in the Barents Sea is controlled primarily by thermodynamic process, and seaice dynamic process plays a very important role in the Greenland Sea; (3) Both the observed and simulated seaice seasonal cycle in the Barents and Greenland Seas lag behind seasonal cycle of surface air temperature, but the lag time of the later is more sigmficant. In fact, it is lagged behind 1～2 months when the simulated sea ice reached its maximum, and there is no sea ice in the Greenland Sea for 2 more months. These are two major aspects in this model which should be improved in the future. 更多还原