Status: Under development, Date: 2025/07/07 10:18, Version: 001

The use of shared, community-endorsed vocabularies for metadata annotation is key to ensuring unambiguous and standardized descriptions of data. This not only supports the alignment and integration of heterogeneous datasets but also enhances data discovery and reuse. Crucially, such practices form the foundation for machine-readability of metadata, which is essential for achieving semantic interoperability.

Within the Helmholtz research field Earth and Environment, there is a growing need for consistent approaches to metadata annotation that ensure semantic interoperability. This recommendation aims to address that need by guiding the selection and prioritization of controlled vocabularies and by supporting the optimization of metadata annotation workflows.

Data infrastructures and data hosts—such as data repositories, sensor registries, electronic lab notebooks, or other platforms that store and make data available—should ensure the annotation of the vast majority of metadata using standardized terms from established and, where appropriate, FAIR-compliant controlled vocabularies (e.g., lists, thesauruses, taxonomies, standardized terminologies, or, ideally, ontologies) to promote semantic consistency, clarity, and interoperability. Data platforms such as data portals or knowledge graphs should incorporate these terms into their tools and reuse them for standardized representation and improved search.

The basis for a comprehensive metadata annotation is the is that data is provided with sufficient and structured metadata and that there is agreement about which metadata is considered essential in communities. Metadata annotation with semantic ressources is only effective if there is consensus within a research community about which controlled vocabularies or other semantic resources best meet the community's needs, and if these resources have clear governance, provenance, and documentation. Furthermore, they should be available and maintained over the long term (at least 5 years) and cover the vast majority of requirements.

Consistent implementation of semantic concepts across a research community is essential for advancing the FAIR principles—particularly Findability, Interoperability, and Reusability. By relying on shared, community-endorsed vocabularies and standardized metadata structures, data can be described in a clear and unambiguous way, enabling automated systems to interpret, link, and integrate datasets across disciplinary and institutional boundaries. This supports not only transparent and reproducible research but also facilitates the meaningful reuse of data. Especially in complex domains such as Earth and Environment, semantic concepts serve as a means to standardize metadata, promote interdisciplinary understanding, and represent complex scientific phenomena in a machine-readable, structured format.

In the current German research landscape, where environmental data are fragmented, inconsistent in quality, and lack standard formats, consistent semantic annotation is key to improving interoperability. Community agreement on essential metadata elements ensures data are well-described, discoverable, and reusable. This recommendation supports the Helmholtz Earth and Environment community by guiding the selection of controlled vocabularies and the improvement of annotation workflows.

What is meant with „controlled vocabulary“ in these recommendations?

There are different types of structured terminologies used in semantic data annotation, each offering varying levels of complexity and expressiveness. According to Le Franc et al. (2019), these can be understood as part of a spectrum of controlled vocabularies, ranging from simple to highly formalized structures. In our recommendations, the term controlled vocabularies is used in this broad sense and thus also includes ontologies.

• A glossary is a simple, alphabetically ordered list of terms from a specific domain, each accompanied by a definition. It supports a shared understanding of terminology within a community.

• A taxonomy is a controlled vocabulary with a hierarchical structure. Terms are related via broader–narrower (parent–child) relationships, and the taxonomy helps classify concepts and organize knowledge systematically.

• A thesaurus builds upon a taxonomy by incorporating not only hierarchical but also associative (e.g., related terms) and equivalence relationships (e.g., synonyms or preferred terms). Thesauri are useful for enhancing semantic navigation and improving information retrieval.

• An ontology represents the most expressive and formal type of controlled vocabulary. While it may include terms and structures from glossaries, taxonomies, or thesauri, it adds formal semantics through logical relationships defined in machine-readable languages (e.g., OWL, Description Logic). Ontologies enable reasoning, inference, and advanced semantic interoperability across systems.

In line with Le Franc et al. (2019), we explicitly consider ontologies to be part of the family of controlled vocabularies addressed in this recommendation. Depending on the use case and required level of formality, different types of controlled vocabularies may be appropriate for metadata annotation and semantic integration.

-

Note: While the benefits of standardized semantic practices are clear, their successful implementation depends on collective coordination, sustainable infrastructure, and adequate funding. Efforts should focus on reusing and adapting existing resources where possible, rather than creating isolated or redundant solutions.

Data stewards, archivists, and tool developers—including those responsible for systems used at various stages of the data lifecycle, such as data acquisition, processing, documentation, and storage—should ensure that metadata is captured in a structured and standardized manner, using harmonized metadata schemas aligned with community standards. This includes platforms such as electronic lab notebooks, archiving tools, sensor registries, and other software environments that support data generation, transformation, or submission. Metadata must be consistently annotated with well-governed controlled vocabularies to guarantee semantic clarity, interoperability, and long-term reusability across diverse data infrastructures. Providing clear documentation of the vocabularies and semantic resources in use, alongside transparent, user-friendly annotation workflows, supports consistent metadata quality and facilitates semantic integration.

Developers of data portals, knowledge graphs, and discovery tools should incorporate these controlled vocabularies and ontologies into their software environments. This enhances machine-readability, promotes semantic consistency across systems, and enables users to efficiently search, filter, and combine data from multiple sources.

To enable seamless semantic annotation from the start, data producers need to be supported through targeted training and awareness initiatives that emphasize the use of community-endorsed vocabularies, structured metadata practices, and annotation best practices. Transparent user guidance and easily accessible documentation of recommended semantic resources are essential to ensure metadata quality and simplify the semantic linkage of data throughout its lifecycle.

- Wilkinson, M. D., Dumontier, M., Aalbersberg, I. J., Appleton, G., Axton, M., Baak, A., … & Mons, B. (2016). [The FAIR Guiding Principles for scientific data management and stewardship](https://doi.org/10.1038/sdata.2016.18). *Scientific Data*, 3, 160018. https://doi.org/10.1038/sdata.2016.18

- Le Franc, Y., Hettne, K., & Ó Carragáin, E. (2019). *D2.5 FAIR Semantics Recommendations Second Iteration*. Zenodo. https://doi.org/10.5281/zenodo.4314321