农业科研试验基地数据管理标准体系构建

农业科学数据作为重要战略性资源,在推动农业科技创新、增强农业科技竞争力方面发挥重要作用。农业科研试验基地是一手科研数据的主要来源,其数据管理标准体系的建立不仅是试验基地数据规范化管理的基础,也是规范化数据管理工作的重要一环,对于提高数据可复用性,最大程度发挥数据价值有重要意义。该研究通过国内外涉农长期观测网络、试验站数据管理标准体系内容梳理,以及对中国农业科研试验基地数据特点、数据全生命周期管理流程调研,构建了农业科研试验基地数据管理统一工作流;基于"霍尔三维结构"从标准适用数据管理阶段、标准性质、标准专业领域3个维度构建农业科研试验基地数据管理标准体系框架,并编制了农业科研试验基地数据管理标准体系基本构成表,梳理出涵盖数据全生命周期管理的标准20项,首期必建标准12项,以期为推进农业科研试验基地建设和数据管理共享提供支撑。

Establishment of data management standard system for agricultural scientific research and experiment station

Abstract: Agricultural science data is an important strategic resource. As first-hand scientific research data producers, agricultural scientific research and experiment station gathers massive scientific research data of different types which are in a loosely managed state. So, the establishment of its data management standard system is not only a basis for standardized data management but also an important part of standardized data management work. And it is of great significance for improving data reusability and fully developing data value. In this research, by investigating the data characteristics and the whole life cycle management process of the agricultural scientific research and experiment station in China, we found that the data had four characteristics: 1) multi-source and isomerism; 2) Small data centralized management, big data decentralized management; 3) short time duration and poor continuity of data; 4) low data standardization and insufficient data sharing. And based on the problem of "small data centralized management, big data decentralized management", we explored the workflow for routine data management in agricultural scientific research and experiment station, then designed a unified workflow. Under the fragmented data management workflow, it was usually a lack of professional and unified standard system in agricultural scientific research and experiment station, so the data sharing and reusability rate were very low, and the data value was hard to realize. But, based on a unified data management standard system, there were the specialized person responsible for data management and sharing and provided services for users, which effectively guaranteed the data quality, greatly improved the probability of the users to discover the data, and reduced the workload of the data owners to manage the data. Additionally, based on the "Hall three-dimensional structure", we constructed the standard system framework for data management of agricultural scientific research and experiment station from three dimensions: standard application data management stage, standard features, and standard professional field. From the dimension of standard application data management stage, the data management standard system should focus on the life-cycle management process such as data collection, data processing, data archiving, data storage and sharing, and each data management stage should have corresponding standards as the basis and guidance. From the dimension of the standard features, the data management standard system should include technical standards, work standards, and management standards. And different types of standards complemented each other to ensure the orderly implementation of data management. From the dimension of the standard professional field, guidance standard was the foundation, data standard was the core and management standard was the guarantee. Based on the previous research and two rounds of expert review and discussion, the data management standard system table of agricultural scientific research and experiment station was compiled. It contained 20 standards covering the whole life cycle of data management of data. Among them, 12 standards were mandatory construction standards for the first construction phase. This table could be used to guide the construction of a personalized standard system for different agricultural scientific research and experiment stations, to achieve the maximum standardization effect with minimum resource input. This standard system was the basis of standardized data management of agricultural scientific research and experiment station and was of great significance to promote the construction of agricultural scientific research and experiment station. However, there is still a big challenge: how to get through the data collection channels to centralize the data (especially the long tail data) from different data producers? A complete collaboration mechanism is also needed.