面向数值预报的地面资料质量控制与同化技术研究

项目名称： 面向数值预报的地面资料质量控制与同化技术研究

项目编号： No.41275105

项目类型： 面上项目

立项/批准年度： 2013

项目学科： 天文学、地球科学

项目作者： 徐枝芳

作者单位： 国家气象中心

项目金额： 80万元

中文摘要： 本项目拟通过双权重、递进式EOF方法及变分质量控制方法分析和建立面向数值同化的地面观测资料质量控制技术，确保参与同化的地面资料的质量满足同化要求，从而提高地面资料同化分析质量和效果；采用边界层相似理论及观测资料统计分析相结合的方式建立地面资料同化算子，利用双权重偏差订正及地形代表误差方法解决地面资料同化中由于地面资料的日变化特征、模式地形与观测站地形高度差异引起的负面影响；结合地面资料同化的要点，通过加强地面资料在边界层内的信息往自由大气的传递解决地面资料同化信息易频散的问题。本项目主要目的是实现我国所有（国家级和省级）地面观测资料在我国短临预报系统GRAPES_RAFS（Rapid Analysis and Forecast System）系统中的合理有效应用，对GRAPES_RAFS系统的分析和短时临近预报有明显正贡献，提高我国地面资料在中尺度数值预报中的使用能力。

中文关键词： 地面资料；同化；质量控制；GRAPES_RAFS；

英文摘要： In this project, bi-weight mean and progressive EOF methods and variational quality control are proposed to address the quality control of surface observation data. The bi-weight mean method is less sensitive to outliers, while the progressive EOF method can not only systematically remove erroneous data and retain good quality data, but also verify if the Gaussian error assumption is approximately satisfied for both the data that passes QC and the background field that is used for assimilation. The assimilation scheme of surface observation data will be devised in this project based on the similarity theory and the statistic analysis results of observation data. The bi-weight mean method and terrain error of representativeness are used to solving the negative influence caused by the different elevation between observation sites and numerical model surface. The bi-weight mean method is also used to correct the different diurnal cycle between surface observation data and the background field. As the information of surface observation data used in assimilation is difficult to be absorbed by numerical model, different measures with different meteorologic element are used to strengthen the information from boundary layer to free atmosphere. The aim of this project is that the surface observation data all over the cou

英文关键词： surface observation data；data assimilation；quality control；GRAPES_RAFS；