使用卫星资料进行边界层四维变分同化研究综述

数值天气预报是当前天气预报的重要手段之一,而同化工作一直是数值预报研究的重点。在查阅了近年来国内外相关文献的基础上,对使用卫星举行资料边界层四维变分同化的研究进展进行了论述,并对最优插值法和变分法2种主要的研究方法进行了简单介绍。研究结果认为,变分方法对于初始场的改善具有明显的效果,与纯粹的统计插值方法相比,变分法具有十分明显的优势;理论上变分法能够同化所有类型大气探测资料,对非常规探测的包容能力极强,不仅能最大限度地获得观测中的信息量,而且避免了各种反演不适定问题。

Overview of Using Satellite Data on Four-Dimensional Variational Data Assimilation on Lateral Layer

At present numerical weather prediction is one of the major methods in weather forecasting, and data assimilation is important. Though referring to domestic and overseas related articles for the past few years, a little comprehensive summarization for the study development using satellite data on later layer with four-dimensional variational data assimilation is presented, and the main research methods in data assimilation-optimal interpolation and variation method are introduced simply. The result shows that variational method makes obvious improvement for initial field, compared with the statistical interpolation method, variational data assimilation has apparent dominance. Theoretically, variation method can assimilate all types of atmospheric detecting data, and having strong assimilation capacity for non-conventional observation data. Not only can it acquire the information in observation to the utmost extent, but also it avoids different kinds of ill-posed problem.