Adapting a Key Semantic Interoperability Innovation from e-Health to Earth Informatics: Are Two-Level Information Models Relevant?

As the current state-of-the-art, ontologies have supported the need for semantic interoperability and data reusability by enabling consistent descriptions of information in the Earth Sciences domain. However, ontologies alone cannot enable semantic interoperability (or data reusability) in the Earth Sciences domain, particularly in the complex modelling of dynamic environments. A challenge lies in integrating heterogeneous Earth Observation data, where high semantic standards, scalability, and flexibility are essential to handle the continuously expanding (and ever-changing) data volume, complex descriptions, and relations of the multifaceted real-world entities, and improve their usability across platforms. Ontologies typically represent comparatively static knowledge which affects their utility in situations like these where modelling dynamic, ever-evolving processes is required on an ongoing basis. They also, at times, lack the necessary descriptive constructs for documenting complex, real-life applications, affecting data interoperability and reusability. In information systems, particularly those in the Earth Sciences, the documentation aspect of defining the concepts, relationships, and terminologies for real-world entities in a dataset is integral to enable semantic interoperability and data reusability. In such continuously evolving environments, Archetypes metadata constructs that can act as high-quality and rigorous documentation templates, exhibit their potential to work together with ontologies to address the challenge of semantic interoperability and maximise the reuse potential of Earth science data. In the archetype-based two-level information models that have been developed over the past twenty years in e-health, the reference model is used to define the core structure and relationships of data while the archetype model is used to provide the domain-specific details separately. A consistent reference model ensures a standardised data format which reduces inconsistencies in data. The archetype model allows definitions that are tailored for a specific domain where archetypes are designed to represent all the necessary information within that domain. The archetypes ensure that all required data fields are collected consistently as defined in the archetype model, ensuring uniform data collection across the domain. The archetypes further help with metadata management by defining data elements and attributes needed for each domain. Within their structure, archetypes also allow the binding of ontologies to provide classification (and associated meaning) of the entities. Over two decades of experience with the use of two-level models for e-health documentation has shown that combining ontologies in an archetype-based two-level information model helps create well-structured and meaningful data and associated information systems. The archetype-based two-level information model approach ensures consistency, enhanced data integration, reusability, and semantic interoperability. This approach, if explored and applied on a domain-wide level, has the potential to make Earth information systems flexible, scalable, more reliable, and efficient in supporting decision-making surrounding environmental sustainability.