Semantic Linking of Research Infrastructure Metadata

The use of metadata to characterise scientific datasets, making data easier to discover and use directly by researchers and via various online data services, is one of the primary concerns of research infrastructures (RIs); also, of concern is the use of metadata to describe equipment, facilities, services and other research assets. Metadata models and terminology differ greatly between different communities and infrastructures, however, and so make synthesising complex interdisciplinary scientific workflows involving as-sets from multiple RIs very challenging.
‘Semantic linking’ addresses the need to enhance the interoperability of RI services and data by bridging metadata schemes, ontologies and vocabularies used by different research communities, whether by standardising the terminologies and schemes used by those communities, or by dynamically transforming metadata from one standard to another when retrieved by services on behalf of researchers executing their scientific workflows.
Multiple techniques for and modes of semantic linking have been investigated in the context of the ENVRI community cluster of environmental and earth science RIs, including top-down modelling of entities and activities within a standard reference model, enrichment of existing metadata records with shared terminology, a full transformation of metadata records from one standard to another, and the generation of additional links to existing online data. We review some of these activities and their application to the promotion of semantic interoperability between RIs, and discuss other possibilities and recent developments that may also be useful for enhancing interdisciplinary data science.