Implementing FAIR Principles for Earth System Data: Insights from the European Eddy-Rich Earth-System Models (EERIE) project

We attempt to make EERIE data FAIR (Findable, Accessible, Interoperable, and Reusable) to enhance its scientific impact and utility. These principles of FAIRness ensure global access, integration, and reuse by researchers and decision-makers, thereby promoting collaboration and innovation.

Findability is enhanced through persistent identifiers such as DOIs and PIDs, ensuring data remains reliably locatable. Metadata standards, including CF conventions and CMIP standard names, ensure precise and efficient searchability. We enhance findability through data catalogs produced in the EERIE and nextGEMS projects, as well as platforms like World Data Center for Climate (WDCC) and DOKU. The WDCC ensures long-term storage with a focus on FAIRness, quality control, and DOI assignment following CF standards. Our EERIE data is also archived on DOKU with PIDs to ensure discoverability.

Accessibility is ensured by providing data through open protocols with clear terms of use. While accessibility does not always mean free access, it guarantees transparency and ease of use. Open-access repositories such as EERIE Cloud, Earth System Grid Federation (ESGF), and, WDCC combination with standardized formats such as NetCDF and Zarr, ensure broad accessibility. Additionally, tools like Zarr provide API access via HTTP, facilitating seamless and efficient data retrieval.

Interoperability is fundamental for integrating datasets across disciplines and platforms. The EERIE project supports this by linking datasets through initiatives such as EERIE Cloud, FREVA and by using standards such as CF conventions to ensure compatibility, facilitating multidisciplinary research.

Reusability is supported through detailed metadata, clear licensing models like CC-BY and CC0, and strong version control practices (e.g, v20240304). Documentation platforms such as easy.gems.dkrz.de assist users to understand and reproduce results. The maintenance of high data quality and the emphasis on archival and replication further enhance the long-term scientific use of these datasets.

Despite these efforts, the implementation of the FAIR data principles in a comprehensive manner poses significant challenges. In the EERIE project, for instance, we work with vast amounts of data, and standardizing it (e.g., CMORizing) can be complex. Obtaining CF-compliant names for all variables is particularly difficult, as there is often no one-to-one documentation from modeling groups. In some cases, this requires manually analyzing code to determine the correct definitions for certain variables.

For climate science, the application of FAIR principles is transformative. These efforts promote global collaboration, enhance the transparency of climate models, and equip policymakers with reliable data to address critical challenges such as climate adaptation and mitigation. Initiatives like EERIE cloud, ESGF and advancements in data processing, such as kerchunking massive datasets, further enhance the FAIRness of climate data, driving innovation and impact.