Argo资料协同管理方法研究

摘要

Argo资料已成为当前海洋和大气科学领域基础研究的重要数据来源.但由于其剖面元数据与观测数据混合存放的特点,现有的共享平台无法实现浮标漂移轨迹与剖面图的实时绘制.因此本文提出一种结构化与半结构化并存的Argo资料协同管理方法,通过分析Argo资料结构组成与特点,将结构化数据与半结构化数据分离提取;然后利用关系型数据库对半结构化类型属性的扩展支持,建立剖面元数据与观测数据间的关联关系;并利用分表存储,降低数据量快速增长对单数据表带来的存储压力.最后通过对近20 a的全球Argo资料解析建库结果进行分析,证明该方法具有良好的可扩展性和高效的轨迹数据获取效率,能够支持浮标漂移轨迹和剖面图的实时绘制.同时,该方法也可为特征类似的剖面观测数据管理提供技术参考.%Argo data have become an important data source in basic researches of ocean and atmosphere sciences .But with the hybrid storage of profile metadata and its observations ,it is difficult for existing shared platforms to realize the real-time visualizing of the drift trajectory and profile chart .A method of collaborative management of structured and semi-structured Argo data were proposed .Based on the analysis of Argo data structure and characteristics ,Argo data were separated as structured and semi-structured data .By means of extended support of relational database to a semi-structured property ,a relationship between profile metadata and observation data was established .Faced of the rapid growth of data ,a multi-table storage method was used to reduce storage pressure of single data table .Finally ,an Argo database was set up and global Argo data in period of 1997-2016 are parsed and imported .The results indicate that the method has good scalability of both storage and query ,as well as high efficiency of float trajectory query ,and it is able to support real-time visualization of float trajectory and profile chart .Our result will offer a reference for management of other similar observation profile data .