全球Argo资料数据管理与共享方法研究

发布时间：2018-05-20 10:19

本文选题：剖面 + Argo　；参考：《浙江大学》2017年硕士论文

【摘要】：Argo计划于1998年被提出至今,已在全球海域范围内投放了约12000个浮标。Argo资料的剖面与轨迹数据对海洋和大气科学领域基础研究具有非常重要的科学意义。国内外相继开发了多种针对Argo资料的共享平台,但在轨迹与剖面图的共享方面,都存在实时性与灵活性差、更新代价高等不足,影响了 Argo资料的获取利用和Argo计划的推广。为了解决以上问题,本文从Argo资料自身结构组成入手,分析了原始Argo剖面资料的结构特性:结构化的元数据存储属性信息,半结构化的实际观测数据以多列数字矩阵的形式存储在文本中。进而提出了结构化与半结构化并存的Argo资料协同管理方法,建立Argo数据库,设计了基于服务的多种数据访问接口,并实现了浮标漂移轨迹的高效获取与实时可视化,重点解决了以下几个关键问题:(1)设计了一种多层树状JSON数据模型,用于组织提取出的剖面观测数据,解决了原始资料不易解析、无法管理的难题,模型能够适应"全球Argo海洋观测网"中不断变化的观测参数、观测深度和采样分辨率的要求,并利用关系型数据库对半结构化属性的扩展支持,建立观测数据与元数据的关联关系,完成了 Argo数据库的构建。(2)针对数据量快速增长的剖面数据,提出了按年份分表的存储方法,有效地降低了单数据表的存储压力,并通过对近20年的全球Argo资料建库结果进行分析,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。(3)实现了多浮标轨迹动态可视化。首先利用线性插值方法,统一多轨迹间的点绘制频率,在通过设置延迟绘制时间,实现了多轨迹异步绘制下的时间同步优化。此外,利用添加辅助线的方法,巧妙地解决了 Web墨卡托投影下跨180度经线轨迹的绘制错误问题。最后,利用Web Service、地图切片与双缓存技术、FSH(Flex + Spring + Hibernate)框架等,建立Argo资料共享服务平台,集成了包括多条件查询、海区查询、多轨迹实时绘制、可交互的剖面图等多种功能模块。该平台已于中国Argo实时资料中心部署并业务化运行,能够快速、方便地为国内外用户提供Argo信息和资料服务,有利于推进Argo资料的共享与使用。研究成果也可为特征类似的剖面观测数据管理提供技术参考
[Abstract]:The Argo plan was put forward in 1998, and the profile and trajectory data of about 12 000 buoys. Argo data have important scientific significance for the basic research in the field of ocean and atmosphere science. Many kinds of Argo data sharing platforms have been developed at home and abroad, but in the aspect of track and profile sharing, the real-time performance and flexibility are poor, and the update cost is high, which affects the access and utilization of Argo data and the promotion of Argo plan. In order to solve the above problems, this paper starts with the structure composition of Argo data, and analyzes the structural characteristics of the original Argo profile data: structured metadata store attribute information, Semi-structured actual observation data is stored in text as a multi-column digital matrix. Furthermore, the co-management method of structured and semi-structured Argo data is put forward, the Argo database is established, a variety of data access interfaces based on services are designed, and the efficient acquisition and real-time visualization of buoy drift trajectory are realized. This paper focuses on solving the following key questions: 1) designed a multi-layer tree JSON data model for organizing the extracted profile observation data, and solved the difficult problem that the original data is difficult to analyze and manage. The model can meet the requirements of changing observation parameters, observation depth and sampling resolution in the Global Argo Ocean observing Network, and use the relational database to extend the semi-structured attributes. The relationship between observation data and metadata is established, and the construction of Argo database is completed. (2) aiming at the rapidly increasing data volume of profile data, a method of storing table by year is put forward, which effectively reduces the storage pressure of single data table. By analyzing the results of global Argo database building in recent 20 years, it is proved that this method has good expansibility and efficient track data acquisition efficiency. It realizes the dynamic visualization of multi-buoy trajectory. Firstly, the linear interpolation method is used to unify the point drawing frequency between multiple trajectories. By setting the delay time, the time synchronization optimization is realized under the condition of asynchronous multi-locus rendering. In addition, by adding auxiliary line, the problem of drawing the trajectory of 180-degree longitude under Web Mercator projection is solved skillfully. Finally, using the framework of Web Service, map slicing and double cache technology, the Argo data sharing service platform is established, which includes multi-condition query, sea area query, multi-track real-time rendering, etc. Interactive profile and other functional modules. The platform has been deployed in China's Argo real-time data center and is operational. It can quickly and conveniently provide Argo information and data services to domestic and foreign users, and is conducive to the sharing and use of Argo data. The research results can also provide a technical reference for the management of profile observation data with similar characteristics.
【学位授予单位】：浙江大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：P715.2;P717

【相似文献】