Analysis and prediction of permafrost change are hampered by lack of observational data. We implement a permafrost data management system to support multi-variable permafrost observation networks. It addresses five key challenges we identified for permafrost data management and publication: (1) existing data management strategies do not scale well, (2) data users have different skills and needs, (3) permafrost data are varied, (4) resources for permafrost data management are limited, and (5) existing permafrost data sources are difficult to integrate. This prototype system supports a workflow from observation to the distribution of interoperable data. It comprises a database with auxiliary software, a set of practices, and an ERDDAP server. We find the system simplifies data publication, hurdles in adapting CF conventions and ERDDAP for permafrost data remain, and technology cannot replace good practice. Our learning can inform organizations who collect, manage, or distribute permafrost data or those who manage large observation networks.

中文翻译：

---

用于多变量永久冻土观测网络的原型现场到发布数据系统

由于缺乏观测数据，永久冻土变化的分析和预测受到阻碍。我们实施永久冻土数据管理系统来支持多变量永久冻土观测网络。它解决了我们在永久冻土数据管理和发布方面发现的五个关键挑战：(1) 现有数据管理策略不能很好地扩展，(2) 数据用户有不同的技能和需求，(3) 永久冻土数据多种多样，(4) 资源多年冻土数据管理有限，(5)现有多年冻土数据源难以整合。该原型系统支持从观察到分发可互操作数据的工作流程。它包括一个带有辅助软件的数据库、一组实践和一个 ERDDAP 服务器。我们发现该系统简化了数据发布，但针对永久冻土数据采用 CF 公约和 ERDDAP 的障碍仍然存在，而且技术无法取代良好实践。我们的学习可以为收集、管理或分发永久冻土数据的组织或管理大型观测网络的组织提供信息。