在分析南海海洋气象观测现状和发展规划的基础上，发现有必要发展合适的资料同化技术，提高资料利用效率，才能克服海洋气象资料相对缺乏来提升数值预报水平。利用多尺度/分块逐批资料同化技术，进行热带气旋资料同化试验。结果表明：多尺度/分块逐批资料同化技术能够适应具有多尺度特征的热带气旋资料同化，较好地满足热带气旋资料同化对流依赖背景误差协方差同化技术的要求，能够较合理高效地利用海洋气象观测资料。基于多尺度/分块逐批资料同化技术，加强海洋气象观测系统的应用，是形成高质量热带气旋初值环流的一种有效途径。

Based on the current situation and development plan of marine meteorological observation, it is recognized that there is a need to develop appropriate data assimilation technology for enhancing the efficiency of data utilization. Only in that way, there is a chance to overcome the lack of observation, and to improve numerical weather prediction. In this paper, the multi scale/block batch wise data assimilation is suggested to perform the test of tropical cyclone data assimilation. The results show: the multi scale/block batch wise data assimilation can be appropriate for the data assimilation of tropical cyclone multi scale circulation, satisfy with the flow dependent background error covariance required by tropical cyclone data assimilation, also can use effectively the marine meteorological observation. By means of the multi scale/block batch wise data assimilation, to amplify the utilization of marine meteorological observation, it is an effective approach to obtain high quality tropical cyclone initial circulation.