The Path to Creating Open Global Networks of All Types of Geochemical Data from Multiple Countries, Repositories and Researchers: what variables and vocabularies are YOU using to describe your data?

Abstract

Geochemical datasets are notoriously heterogeneous and are collected by thousands of researchers/research groups on a diversity of samples (rocks, minerals, meteorites, fluids, gases, etc) using hundreds of analytical techniques across multiple disciplines around the globe. Even where there is consensus on the minimum variables needed to be reported for a particular analytical method, the vocabularies used for each variable are rarely consistent, making automated integration of equivalent data types extremely difficult.

New, data-driven research questions are emerging that require globally aggregated geochemical datasets to enhance our understanding of planetary system processes, for which AI and ML tools are becoming increasingly employed. Combined, these applications are exacerbating the need to make geochemical data Findable, Accessible, Interoperable and Reusable (FAIR) thus driving a demand for standardisation of geochemical variables and vocabularies. Unfortunately there is presently little coordination and multiple vocabularies are emerging that often replicate the same concepts.

The OneGeochemistry Initiative (http://onegeochemistry.org/) is working towards enabling a Open Global Network of all types of geochemical data from multiple countries, repositories and researchers. To achieve a balance between meeting increasing demands for local FAIR vocabularies versus time required to reach international agreement, OneGeochemistry developed a three-tiered approach (local, community, international):

1) Encouraging data providers with locally defined vocabularies to make them available online and ensure each term has a persistent identifier;

2) Encouraging groups with similar topics to begin harmonising on multiple locally-derived concepts/definitions and publish these as community resources;

3) Raising awareness of groups harmonising and making semantic resources FAIR-compliant at an international level, particularly those with endorsement from International authoritative groups (eg Scientific Unions/Societies).

As convergence takes place towards internationally-agreed terms, the size of the community able to share (meta)data machine to machine grows. As better defined vocabularies become available they should be easy to translate into multiple languages.