Project HF-EOLUS. Task 2. VILA & PRIO Radial Metrics GeoParquet Dataset.

This record publishes the outputs of the HF‑EOLUS HF‑radar radial metrics ingestion pipeline, produced with the HF‑EOLUS ingestion code (https://doi.org/10.5281/zenodo.17071947), as a single archive: hf_eolus_hf_ingestion.tar.gz. It contains a STAC 1.1 catalog for stations VILA and PRIO (collections HF‑Radar‑VILA1..4 and HF‑Radar‑PRIO1..2; VILA: 2011‑09–2023‑05; PRIO: 2011‑08–2015‑04 and 2018‑05–2023‑11) and GeoParquet v1.1 datasets for radial metrics (radial_metrics/), header (header/), and range summaries (rng_info/).

The LLUV inputs used for ingestion are archived at https://doi.org/10.5281/zenodo.16528652 and https://doi.org/10.5281/zenodo.16458693. First, these LLUV files were uploaded to AWS S3; the included pipeline.sh then orchestrates their processing: AWS Glue parses LLUV into partitioned Parquet, Athena registers external tables, and STAC JSON is generated. The STAC property templates applied during generation, stac_properties_collection.json and stac_properties_item.json, are bundled for transparency and reuse.

After unpacking, open hf_ingestion/catalog.json with a STAC client to browse collections and items. Parquet assets are partitioned by pos_bragg and hourly timestamp; geometries are WKB points in WGS84 stored in the geometry column (example fields include Pwr and VELO).

Total number of rows (radar echoes) in the radial_metrics datasets:

• VILA: 347,275,384
• PRIO: 240,847,334

Acknowledgements