Developing a Blueprint for Open Science in the AuScope Research Infrastructure by Enabling Vertical Integration Across Multiple Levels of Processing.

Open Science mandates set a high bar for reproducibility and transparency, requiring open knowledge of all input/output sample and data artefacts. It also requires identification of any actor and tools used to process and model these along the full Research Workflow, starting from acquisition of the Primary Observation Datasets (PODs) and samples, through initial data calibration and then subsequent generations of subsamples and digital outputs. At the same time, compliance with the FAIR and CARE principles, and demands for AI-Ready and Decision-Ready data are also ubiquitous: compliance with all these across multiple levels of processing adds additional complexity to the Open Science Paradigm.

AuScope is Australia’s national geoscience Research Infrastructure (RI) funded through the National Collaborative Research Infrastructure Strategy (NCRIS). AuScope facilities enable the collection of multiple data types, ranging from drone, geophysical and satellite data collections that can be TBs in volume, down to small-scale (MB) long tail collections in geochemistry and geochronology. As articulated in the AuScope Research Data Systems Strategy 2025-2030 (https://doi.org/10.5281/zenodo.15825498), AuScope is committed to Open Science and ensuring compliance with FAIR and CARE. 

Using examples from two AuScope Opportunity Fund projects in geophysics and geochemistry, this paper demonstrates how AuScope is developing a Blueprint for a methodology for Open Science that also establishes compliance with FAIR, CARE, AI-Ready and Decision-Ready requirements at each level of processing. Where possible, all research input and output artifacts are Findable, Accessible, Interoperable and Reusable by machines and CARE is being implemented to complement FAIR and document Indigenous interests and governance. It is planned for AI-Ready data to build on FAIR and CARE with additional metadata on quality, documentation, access and preparation, whilst Decision-Ready guidelines will be implemented through a chain-of-custody approach that allows tracking of any activity, any actor involved and any transformation undertaken.

The first stage in the Blueprint is that for each data type, definitions are agreed for the various levels of processing, starting with PODs through to derivative data products, models and visualisations. Where possible, the NASA processing levels are followed: however, more specific definitions have been created for geochemistry, hyperspectral and magnetotelluric data, with additional definitions planned for other data types. 

Identifiers such as ORCIDs, RORs, RAiD, IGSNs and DOIs are essential at each level of processing to uniquely identify the contributing researchers, research infrastructures, funders, software developers, software etc., and allow connections across each successive level of processing. Identifiers will 1) enhance ways for credit to be given to researchers and funders at each processing level and 2) ensure Indigenous metadata are recorded for each POD and then carried downstream to any derivative product. 

Preliminary results from these AuScope Opportunity Fund projects show that although implementing transparent Open Science is complex, already in geophysics it is allowing researchers to access data at the processing levels most suitable to their research objectives. Ultimately, it is hoped that the transparency enabled in each processing level can contribute to greater trust in solutions proposed for global environmental and social challenges as outlined in the UN Sustainable Development Goals.