Diving Deep into the Murky Waters of Oceanographic Dataset Traceability

Abstract:

Ocean Networks Canada operates observatories and platforms in coastal, deep-ocean and polar environments. Over the last decade, the data repository has become increasingly diverse in terms of data sources, services and partnerships. Ensuring traceability of the data offerings has been a challenge, but also a priority. These records are critical to vetting and understanding the data, both for end-users as well as repository staff.

While details vary from one dataset to another, end-to-end workflows are defined to capture an instrument throughout its life cycle. Transparency is demonstrated by publishing the rawest data, in addition to derived products, although additional work is needed to expose details of the processing chain. Changes are captured with timestamps and responsible person or automated process. The majority of datasets are associated with expedition records such as imagery and reports; while some records are not public, users can often access annotated underwater video from maintenance expeditions at our cabled observatories.

More recently, these dynamic datasets are assigned DataCite Digital Object Identifiers (DOIs). Local query identifiers specify subsets of data requested through our data services, so that users can precisely cite data in terms of time range, product and parameter choices. Dataset versioning results in a new DOI that is associated with its predecessor, using a framework that can capture and display provenance details. Attributions to organizational data partners are included in the DataCite and ISO 19115 metadata, including Research Organization Registry identifiers when available.

Along the way, there have been decision making challenges such as feature prioritization, effort, sustainability, and heterogeneity. Going forward, plans to enhance our provenance and attribution records include more comprehensive ISO 19115 lineage metadata, notification systems, and seamless aggregation and representation of provenance information.