LinkedEarth: a research ecosystem for open and reproducible paleoclimatology

Résumé fourni par la source

Paleoclimate research increasingly relies on the synthesis of heterogeneous proxy records to investigate climate variability across spatial and temporal scales. However, the reuse and integration of paleoclimate data remain limited by inconsistent formats, fragmented analytical workflows, divergent nomenclature and practices, and implicit methodological assumptions. Improved open data access, while helpful, is insufficient to ensure reproducibility and enable large-scale scientific discovery on its own. Here, we present an integrated framework for advancing open and reproducible paleoclimatology through three complementary developments. First, we build on community standards to harmonize the representation of paleoclimate data and metadata, connecting structured data formats with controlled vocabularies via an ontology. Second, we implement these semantic relationships within a knowledge graph, which integrates data and metadata into a unified, queryable structure that supports consistent access across heterogeneous datasets. Third, we introduce a curated gallery of reproducible and executable workflows that document complete analytical pipelines from data discovery to interpretation. Representative examples illustrate how these three pillars can be combined to enable transparent, efficient, and scalable paleoclimate analyses. This integrated approach provides the infrastructure necessary to support large-scale synthesis and reproducible research, laying the foundation for both human- and machine-assisted scientific discovery in paleoclimatology.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.