OBIS-SEAMAP Dataset Versioning & DOI Policy

OBIS-SEAMAP curates and distributes marine species data from individual researchers and external repositories. To ensure clarity, reproducibility, and sustainability, we use the following versioning and DOI policy:

DOIs

A DOI represents the identity of a dataset.

  • Each dataset in OBIS-SEAMAP is assigned a single persistent DOI that links to the most current version of the dataset.
  • This DOI remains stable across updates unless there is a major structural or scientific change warranting a new dataset identity.
  • Users should cite the DOI along with the version number and access date.

Versioning

  • Dataset versions are tracked in a public changelog that lists update history, data changes, and version labels (e.g., v1.0, v1.1).
  • Updates are categorized as:
    • Initial – First release
    • Update – Minor changes to records or metadata (e.g. New records appended, minor metadata additions)
    • Major – Substantial additions, schema changes, or redefinitions
    • Corrective – Fixes to previously published data (e.g. typos in metadata fixed, incorrect data fixed)
  • Dataset version has three parts MAJOR.MINOR.PATCH (e.g. "1.0.0"" or "1.2.1")and the major part increases with a major update. The minor part increases with an update on data or metadata and the patch part increases with a corrective fix.
  • Changelog is included in the metadata.

Update Cycles

  • Datasets from external repositories are automatically harvested and refreshed on a monthly basis. The version number is incremented while the DOI remains the same.
  • Datasets from individual providers (e.g., NOAA researchers) are updated on an as-needed basis, based on provider submissions.

Archived Versions

  • Historical versions are retained internally for reproducibility.
  • Specific versions may be linked on the dataset page where needed.

When DOIs Change

A new DOI may be assigned when:

  • The dataset scope changes significantly (e.g., new species, regions, or methods).
  • The structure of the dataset is redesigned (e.g., new schema, major data integration).
  • A dataset is split or merged with others to form a new, distinct product.

The original dataset(s) may be unpublished, in which case the DOI link for the obsoleted dataset will note this status change.

Slideshow Main Title