Conceptualizing and Assessing the FAIRness of Environmental Scientific Data in Trustworthy Digital Repositories

Abstract:

The lack of interoperability of data archives has been a significant obstacle to interdisciplinary cooperation within environmental sciences. While there has been much progress in developing best practices and standards within designated communities, there is still a great challenge to combine data from different fields. Research data should be FAIR within and across domains, and be organized appropriately for subsequent reuse to address scientific challenges such as geohazard, biodiversity loss and climate change through data-driven solutions.

The uptake of the FAIR principles requires practical solutions (e.g., recommendations, training, and tools) that facilitate their application throughout the research data life cycle. In particular, metrics to assess FAIR digital objects and tools to estimate data FAIRness over time are important. The FAIRsFAIR project contributes to these aspects as part of building the European Open Science Cloud. Existing FAIR models focus on ‘what’ is to be evaluated, i.e., indicators to measure data FAIRness. A gap remains in FAIR applications, i.e., ‘how’ the indicators can be implemented in practice, e.g., to help FAIR stakeholders such as research communities and data service providers improve their data. FAIRsFAIR addressed this issue through the development of fifteen metrics addressing the FAIR principles and a tool to support an automated FAIR assessment of scientific data from Trustworthy Digital Repositories. The metrics are based on indicators from the RDA FAIR Data Maturity Model Working Group and previous projects. We present the metrics and the tool’s development and their iterative evaluations with selected environmental science repositories. The assessment is mainly built on domain-agnostic meta(data) characteristics. Because FAIR data assessment depends on the repository context, we also gathered other criteria related to enabling data reuse as part of the feedback process. The criteria missing from the metrics include those addressing technical quality checks applied by the repositories as part of their operational workflows. Planned work includes extending the assessment by selecting some of these metrics against which practical tests will be implemented.