Nixon A1, Deng R2, Kallas L3, Luzi-Helbing M4, Klöcking M5, Sircombe K6, Lenhert K2, Ware B7, Welti N8, Wyborn L9,10
1Adelaide University, Department of Earth Science, Adelaide, Australia, 2Columbia University, Lamont-Doherty Earth Observatory, Palisades, USA, 3DIGIS, University of Göttingen, Göttingen, Germany, 4GFZ Helmholtz Centre for Geosciences, Potsdam, Germany, 5University of Münster, Institute for Mineralogy, Münster, Germany, 6Geoscience Australia, Energy and Groundwater Division, Canberra, Australia, 7Curtin University, John de Laeter Centre, Perth, Australia, 8CSIRO, Agriculture and Food, Waite, Australia, 9National Computational Infrastructure, Australian National University, Canberra, Australia, 10AuScope Ltd., Melbourne, Australia
Biography:
Angus Nixon completed a PhD in Geology studying the low-temperature evolution of northern Australia and the relationship between plate boundary tectonics and structural and chemical signatures in the continental interior. He is currently an EarthBank Fellow at Adelaide University and part of the AuScope Geochemistry Network, and co-chair of the OneGeochemistry initiative coordinating efforts between international geochemistry data systems. His work includes designing standards for geochemical methods including fission track thermochronology and beta-decay geochronology, as well as resources for describing analytical practices and enable lab-to-repository and machine-to-repository connections.
Abstract:
To address more complex research questions, multiple types of geochemical data/data products are generated by a wide variety of analytical methods and applications. Whilst still imperfect, domain repository options exist at various scales internationally for storing and curating geochemical data results. However, the diverse geochemical data types being generated, combined with Open Science, FAIR, and AI-ready mandates from funders, now require provenance information on the methods used to acquire and reduce data, as this is crucial to providing context for the interpretation of reported measurements. Without these, many geochemical data types cannot be reused with confidence and ensure that like is being compared with like. Unfortunately, there are few community-agreed conventions, vocabularies, protocols, etc., available to report these in standardised, machine-interpretable ways that can be linked as metadata to each analysis.
The OneGeochemistry initiative, comprising eight major geochemical data infrastructures worldwide, is actively developing resources to create machine-interpretable descriptions of the techniques, methods, protocols, and procedures used to acquire geochemical data. The crucial first steps are to systematically define stages of any analytical workflow for each data type and then construct resources that describe the key information at each stage. In an experimental approach, LLMs were used to review literature and help generate recommended best practices on protocol metadata documentation.
This presentation will highlight resources produced so far, including an extensive vocabulary and definitions for the breadth of analytical techniques used across geochemistry, as well as metadata frameworks for analytical protocols that describe data context and quality.
