This innovative project carries out exploratory research applying semantic technologies to support data representation, discovery, sharing, and integration between disparate geoscience data types and structures. It is a risky, high pay-off activity that, if successful, has the potential to transform our ability to discover, access, and use geoscience data in ways not possible at present. The goal of this research is to develop a prototype involving the data collections of some major NSF-funded Data Management Centers: IEDA and R2R at the Lamont Doherty Earth Observatory at Columbia University and BCO-DMO at the Woods Hole Oceanographic Institution. The effort is focused on making NSF-collected data for the ocean sciences and other associated datasets more easily and widely accessible and available to researchers and the public. Linked Open Data methodologies will be employed. Essential elements of this approach include the use of unique data identifiers to mark, link-to, and dereference specific data and details. Once the intial relationships are aligned, the new system can automatically infer new relationships between data and data locations. Goals will be to semantically integrate the data already available from the initially targeted data reposotiries in such a way that the approach can be scaled up to the whole of EarthCube, a new NSF initiative to develop a geoscience knowledge and data mangement system for the 21st Century. This project is a collaboration between ocean science resaerchers, computer scientists, and ocean data management centers from Maryland, Ohio, New York, and Massachusetts. Braoder impacts of the work include building infrastructure for science and improving public accessiblity to NSF-funded data collections.
The overarching broad goal of the one-year collaborative OceanLink project was to bring together ocean science researchers, librarians, data managers and cyberinfrastructure scientists to explore the possibility of using Linked Data to enable the integration of ocean science data repositories, library holdings, conference abstracts, and NSF funded research awards. The partners in this collaborative project sought to develop technical solutions for key issues in the areas of data discovery, sharing and integration for ocean science research data and library holdings. The solutions provide a prototype framework that will aid automated integration of Geoscience datasets as well as provide best practices for semantic integration within the oceanographic subdomain of the EarthCube community. While benefiting the ocean sciences community immediately, the framework and best practices will be relevant to all Geoscience domains represented in the EarthCube research community. The project employed an emerging semantic methodology called Ontology Design Patterns (ODPs) to develop a set of patterns, or partial ontologies, that formalize key notions in ocean science data. These ODPs enable horizontal integration between repositories with potentially independent semantic models by not forcing each individual implementation to conform to a model that does not address their specific needs. In the case of this project, the partners came from related but varying aspects of ocean sciences. These partners were able to map their content to all relevant ODPs without making semantic commitments often required by larger, monolithic ontologies. Integration across all partner content was then made by searching partner content for use of the ODPs. The ODPs are publicly available at http://schema.oceanlink.org/ and will continue to grow and evolve as new concepts emerge in the ocean sciences and other geoscience domains. Furthermore, the growth of the ODPs developed under this project does not conflict with this original work nor does it cause error to existing integrations of ocean science data. Growth, extension of existing ODPs and addition of new ones, provides newer integration opportunities as they are adopted.
