Scaling FAIR Data Practices in Climate Modelling

Making research data Findable, Accessible, Interoperable and Reusable (FAIR) is now widely recognised as essential for open and reproducible science. In practice, however, translating FAIR principles into everyday data management remains challenging, particularly in climate modelling, which involves large data volumes and complex software and data environments on high-performance computing (HPC) platforms. Research rarely follows a simple path from data generation to publication, and FAIR is still often treated as a final, optional step rather than as a set of practices embedded and maintained throughout scientific workflows. 

We present a case study from Australia’s Climate Simulator (ACCESS-NRI) that examines how FAIR principles can be advanced through two complementary approaches applied in parallel. One focuses on the social and practical aspects of FAIR, supporting researchers to apply FAIR practices as part of their everyday research activities. The other centres on embedding FAIR directly into tools and processes, thereby reducing reliance on manual effort and helping to minimise the errors and inconsistencies that naturally arise in complex, collaborative environments. 

Through an open, merit-allocation based approach, ACCESS-NRI provides multiple data sharing pathways, from shorter-term spaces that support active development and collaboration to more curated, publication-ready datasets for longer-term access. This staged model supports the progressive application and uplift of FAIR practices as data are generated, shared, and refined over time, substantially streamlining later curation. Alongside this, we have also focused on improving the consistency and standardisation of ACCESS model outputs by embedding established community conventions and defined data specifications directly in the ACCESS software and release processes. This helps reduce variation across model outputs, supports reuse across tools and researchers, and shifts FAIR from a largely manual effort towards standard practice. 

This case study demonstrates how FAIR principles can be advanced through practical, community-aligned approaches that fit within real research contexts. For ACCESS-NRI, these efforts provide a foundation for tackling deeper FAIR data challenges, with lessons that are relevant to other Earth and environmental science domains facing similar constraints.