Autonomous LLM-agent harmonization of environmental data across global research infrastructures (GERI): input data, metadata, human reference harmonization, target schema, and agent decision logs

This collection is the supplementary research object for a study in which large language model (LLM) agents autonomously harmonize environmental data across five global research infrastructures (NEON, ICOS, TERN, SAEON, and eLTER). It contains five components: (A) the exact raw input-data snapshot the agents consumed; (B) the source metadata (sensor positions, variable dictionaries, site tables, NetCDF attributes); (C) the human-expert reference harmonized dataset (March and May 2026 versions) used to validate the agent output; (D) the a-priori GERI target schema; (E) the per-run agent decision and quality-control logs; (F) the validation matrix data. Per-dataset citations for all 121 source datasets are in A_source_data_citations.csv, and 00_README.md describes the full layout and how to reuse each component. GERI-produced components are released under CC0 1.0; redistributed source datasets retain their own licenses (see the citations file).

Data and Resources

Additional Info

Field Value
Author Karns, Tyler D., Hagen, Cedric J., Deshpande, Krutika, SanClements, Michael D., Laney, Christine, Ruddell, Benjamin L., Loescher, Henry W., Swetnam, Tyson L.
Last Updated September 18, 2026, 03:00 (UTC)
Published August 5, 2026, 18:00 (UTC)
Citation Karns, Tyler D., Hagen, Cedric J., Deshpande, Krutika, SanClements, Michael D., Laney, Christine, Ruddell, Benjamin L., Loescher, Henry W., Swetnam, Tyson L. 2026. Autonomous LLM-agent harmonization of environmental data across global research infrastructures (GERI): input data, metadata, human reference harmonization, target schema, and agent decision logs. CyVerse Data Commons. DOI 10.25739/e9d8-ee77
contributor Stefaniak, Elisa, NEON, ICOS, TERN, SAEON, eLTER
de_created_date 2026-08-25T12:37:52Z
de_modified_date 2026-09-18T02:41:31Z
fundingReference National Science Foundation (NSF), https://doi.org/10.13039/100000001
publisher CyVerse Data Commons
relatedIdentifier https://github.com/GERI-Drought-Pilot/geri-ai-harmonization, https://doi.org/10.5194/egusphere-egu26-12357
resourceType Data harmonization collection
reuse_or_citation_conditions Cite this collection by its DOI. Source datasets must additionally be cited per A_source_data_citations.csv (each carries its own DOI/PID and license). TERN citations are provisional pending finalization with TERN; eLTER source citations are partial