The European Plate Observing System (EPOS) is a European Research Infrastructure Consortium (ERIC) and recognised as a European Strategy Forum on Research Infrastructures (ESFRI) Landmark in solid Earth sciences. EPOS is dedicated to advancing scientific research by integrating and facilitating access to ​data and services generated by ​diverse and distributed research infrastructures across Europe. Its mission is to enable ​excellence science and innovation for ​the monitoring and understanding of the dynamic and complex Earth system by leveraging e-science opportunities and fostering interoperability across multiple scientific disciplines. 

The need for a common metadata profile

A core technical challenge in EPOS was the integration of multidisciplinary data of EPOS’ Thematic Core Services, which are disciplinary infrastructures contributing domain-specific datasets, services, and tools. Each Thematic ​Core ​Service operates independently, adopting different technologies, data formats, metadata standards, and vocabularies, making seamless integration a complex task. EPOS developed a central hub underpinning an ​integration platform​ to address this integration challenge. This unified e-infrastructure aggregates and provides centralised access to heterogeneous assets from various thematic communities.  

Given the variability in technologies and standards, a common metadata application profile was essential to standardising metadata exchange, facilitating data discoverability, and enabling seamless integration of diverse data sources within the EPOS Research Infrastructure. A decision to adopt DCAT-AP was made over other alternatives, which was guided by the need for a flexible and extensible solution that could effectively describe the diverse and heterogeneous datasets, services, and resources provided by various scientific communities.  

Tailoring DCAT-AP for EPOS Needs 

EPOS spans multiple scientific domains, each following distinct metadata conventions, data formats, standards, and vocabularies. DCAT-AP acted as a reference application profile that could be tailored to specific needs, allowing EPOS to accommodate domain-specific metadata requirements. By developing EPOS-DCAT-AP, EPOS was able to customise the profile to meet its specific needs, ensuring that metadata is harmonised across multiple disciplines while preserving its richness and specificity. The sustainability and long-term viability of DCAT-AP were crucial factors in its adoption. EPOS is designed as a long-term infrastructure, which necessitates the use of a metadata model that is well-documented, widely adopted, and actively maintained to ensure its continued relevance. 
 

EPOS began developing EPOS-DCAT-AP in 2017, building the extension on DCAT-AP v1.1, which was primarily designed for datasets shared as static files or through simple download mechanisms. At that time, DCAT-AP did not yet support web services, which posed a significant challenge for EPOS. EPOS encountered several challenges on their way to successfully implement DCAT-AP, requiring innovative solutions to ensure the profile met their needs. 

• Service-based interactions: The EPOS infrastructure relies heavily on service-based interactions to provide access to datasets, software, and research data products. To address this limitation, EPOS extended DCAT-AP by introducing metadata elements for representing services, their endpoints, parameters, and relationships with datasets. With the release of DCAT-AP version 2 and 3, the specification introduced the concept of DataService, improving its capability to describe web-based access points. 
• Lack of real-world examples: When EPOS started working with DCAT-AP, there were few practical references demonstrating how to apply the profile to describe different types of resources. This required experimentation with different serialisation approaches to ensure the implementation remained functional, interoperable, and aligned with broader European data-sharing initiatives. 
• SHACL validation: There was limited guidance on SHACL validation, making it difficult to define and enforce constraints on metadata descriptions. The lack of mature SHACL validation tools meant that ensuring EPOS-DCAT-AP metadata's correctness, consistency, and compliance required manual efforts and iterative refinements. 
• Alignment with evolving specifications: As new versions of DCAT-AP were released, maintaining alignment with evolving specifications became another challenge. Newer versions introduced improvements such as DataService, ensuring compatibility between EPOS-DCAT-AP and successive DCAT-AP updates, which required continuous adaptation and revision. This involved monitoring changes to the standard, adjusting the extension where necessary, and ensuring ongoing interoperability within the EPOS framework. 

Reusing DCAT-AP provided several advantages to EPOS, particularly in accelerating development, ensuring metadata consistency, enabling machine-readability, and fostering integration with external platforms.  

• Accelerated development: Building upon an existing, well-defined metadata model significantly reduced development time and effort, allowing EPOS to focus on extending the model for EPOS-specific needs.
• Metadata consistency: The availability of a common vocabulary and shared semantics simplified metadata management across diverse stakeholders in EPOS. Adopting a common reference model helped establish a more consistent and harmonised metadata profile.
• Machine-readability: The machine-readability of DCAT-AP metadata facilitates automated processing, discovery, and integration of EPOS resources. Since DCAT-AP is RDF-based, it supports semantic web technologies and linked data principles, making metadata more interoperable with external platforms. Ensuring that EPOS metadata can be easily harvested, indexed, and processed by machines, enabling more efficient data discovery and integration with other research infrastructures.
• Ongoing improvements: As DCAT-AP evolves, EPOS-DCAT-AP benefits from continuous improvements and community-driven updates, ensuring long-term sustainability. EPOS metadata remains adaptable to future interoperability requirements, making the system more resilient to technological changes.
• Broader adoption: Adopting a widely used metadata profile encouraged greater adoption and reuse of EPOS metadata beyond its research community. Structuring EPOS metadata according to DCAT-AP increases the likelihood that other infrastructures, research projects, and developers will engage with EPOS data and services, enabling broader scientific collaboration and innovation. 

A concrete example of the impact of EPOS-DCAT-AP is its adoption by other research infrastructures, demonstrating its interoperability and scalability. It has been successfully ​​tested ​within the European Multidisciplinary Seafloor and Water Column Observatory (EMSO) Research Infrastructure as part of the Geo-INQUIRE project. Furthermore, EPOS-DCAT-AP has influenced broader European research initiatives, such as ENVRI-FAIR, ENVRI-HUB-NEXT, and DT-GEO, where it serves as a blueprint for metadata harmonisation across environmental and Earth science infrastructures. 

EPOS successfully developed the EPOS Platform, a metadata-driven system that provides integrated access to solid Earth science datasets and services. EPOS-DCAT-AP has been able to harmonise and collect metadata descriptions for over 300 data services ​currently ​spanning ten disciplines, making it easier for researchers and stakeholders to find, access, and reuse scientific data. The EPOS Platform offers advanced search capabilities, interactive visualisation tools, and harmonised access to distributed data and services, fostering interdisciplinary collaboration and scientific innovation. To encourage further adoption and development, the entire platform has been released as open-source software under the GPLv3 licence, allowing developers and research communities to extend and adapt its capabilities based on their specific needs. 

For more information, please visit EPOS-DCAT-AP GitHub and EPOS website. 

Written jointly by the SEMIC Team and the editors of the EPOS-DCAT-AP.