The Biological and Chemical Oceanography and Data Management Office (BCO-DMO) is an open access digital data repository created to assemble, curate, and publicly serve data and associated products from NSF-funded and related oceanographic research. The project contributes to geoscience research infrastructure by assisting researchers in sharing their project output, educating them on data management best practices, and curating a catalog of high-quality data for use in research, policy, and education. During the 2014-2020 award period, the project successfully transitioned three PIs, during which time it processed and published over 3,300 datasets of diverse types from over 1,000 funded projects, submitted by over 1,400 contributors from physical, chemical, ecological and biogeochemical sub-domains. In addition to core operations, the project leveraged synergistic collaborations (e.g., NSF EarthCube program), to explore and adopt technologies and practices that benefit the scientific user community. In 2017, the office participated in its first mid-term review and subsequently took action to address review recommendations. This prompted an assessment of existing infrastructure, and development of strategies and roadmaps for improving or redesigning several system components. Outcomes described here improve repository infrastructure; data discovery and access; and data management literacy. The project developed resources to lower barriers to data sharing and engage researchers in data management. For example, in 2014, the project created a Data Management Plan (DMP) template to assist researchers in fulfilling proposal requirements and to better anticipate data management needs. The office also constituted a standing advisory committee in 2018 to ensure development of strategies remain responsive to the needs of the oceanographic research community. Community outreach was conducted at several scientific meetings to gather user needs. This informed publication of resources such as the BCO-DMO Quick Guide to aid researchers in submitting their project data, and tutorials on new tool use. Formalized data science training curricula was administered to increase data management literacy within the oceanographic community. The project also collaborated with academic institutions where the BCO-DMO is leveraged in curricula. BCO-DMO also collaborated to share data for use in marine microbial ecology research, where research requires access to environmental data curated separately from genetic information, by providing access to its data through formalized pipelines that are reusable for other domains. The project also contributed to development and adoption of technologies and best practices through working groups, synergistic projects, and volunteer service within the Research Data Alliance, AGU, EarthCube, and Earth Science Information Partners. In evaluating existing software, we identified a need for a more flexible way to move data through submission, processing and archive. To address this, we adopted a method for packaging data and metadata together that allows all BCO-DMO software a shared, common way of accessing data. This strategy dovetailed with our newly developed and implemented data processing tool which improves transportation of data to, from, and within BCO-DMO systems; supports transparency and traceability of activities; and improves efficiency and consistency of data processing. BCO-DMO implemented NOAA’s ERDDAP tool for improved visualizing, subsetting and downloading data in various formats. For mitigating costly future software replacements, BCO-DMO migrated knowledge about its data into a knowledge graph. This technique turns traditional software code into data that can be preserved, altered, and extended with emerging community needs without incurring exorbitant costs. During this award period, we identified a need to update our website, the primary discovery and access point of the BCO-DMO data catalog. Using guidance from NSF and our user community, we began a full website redesign to improve discovery, content delivery, and access of our data holdings. to be implemented during the next award period). We also began tracking analytics on activity across the BCO-DMO data catalog website. To facilitate data publication and researcher attribution, the project implemented use of dataset Digital Object Identifiers (DOIs). Metadata associated with datasets and investigators were expanded to include persistent identifiers for people, funding sources, related publications, and maps that detail data geospatial extent and sampling locations. These additions enable distributed, interoperable data systems to exchange data and information and provide improved context for data interpretation and reuse that foster next-generation research in all disciplines. Responding to recommendations for more accurate search capability, the project implemented open source search software to improve search result, and publishing metadata (schema.org) on its website to increase the discoverability of BCO-DMO data on the web(e.g., by Google). These activities bring the BCO-DMO repository in alignment with current data science best practices, and take advantage of state-of-the-art technologies that allow the repository to remain agile and adapt to future needs of the oceanographic research community. The resulting infrastructure and services help lower the barriers to data sharing and improve the interoperability, discovery, and access of opceanographic data; facilitating integration of diverse datasets to enable researchers to achieve a deeper understanding of ocean ecological and biogeochemical systems. Last Modified: 12/30/2020 Submitted by: Danie Kinkade
©2020 Biological and Chemical Oceanography Data Management Office.