中文摘要：

在简单介绍GPS／LEO掩星探测大气的发展历史和科学意义之后，详细阐述了反演的基本原理；分析了标准反演中存在的问题，并说明一维变分同化（1DVAR）在反演方法中的重要性；给出了一维变分同化中价值函数的求解，以及各种同化因子；简单介绍了对当前气象学中普遍使用的四维变分同化（4DVAR）；重点讨论了各种同化方法，以及使用各种同化因子的优缺点。最后，通过CHAMP卫星的观测实例分析，验证了GPS数据在数值天气预报（NWP）中的作用，以及相对于标准反演法一维变分对气象要素的改进。

英文摘要：

After introducing the history and development of GPS/LEO radio occultation technique, we discuss the theory of retrieval technique. The importance of the 1DVAR is demonstrated in the retrieval technique by analyzing the problems in standard retrieval technique. Then we describe the cost function of 1DVAR and the sorts of assimilation parameters in detail. Furthermore, we introduce the 4DVAR, which is popular used in meteorology nowadays. We especially analyze the different assimilation techniques and the benefits and defects of using different assimilation parameters. At last, by CHAMP radio occultation observation, we prove that the GPS measurement plays an important role in NWP, and compared to the standard retrieval technique, there are also notable improvements after using the 1DVAR.