CalCOFI public storage

3 public buckets · served via https://storage.calcofi.io

bucketcontents
calcofi-db
Integrated database releases		Versioned Parquet releases of the integrated CalCOFI database — observations, samples, taxonomy and reference tables, plus machine-readable catalog/metadata sidecars for each release.
calcofi-files-public
Source archive for reproducible ingests		Primarily a versioned archive of the INPUT files the database ingests read — largely CSVs synced from Google Drive — kept so any release can be rebuilt reproducibly from the exact inputs it was built from, rather than from whatever Drive holds today. Also carries published derived products, such as the whole-dataset CF NetCDF files.
calcofi-projects
Project outputs		Outputs from individual CalCOFI projects and analyses.
Reading these files directly. Everything here is a plain HTTPS object, so most tools can read it in place without downloading. For the Parquet releases:
SELECT * FROM read_parquet('https://storage.calcofi.io/calcofi-db/ducklake/releases/latest-version/parquet/obs.parquet') LIMIT 10;

Resolve latest-version from latest.txt, which always holds the promoted release. Prefer a form? Try the browser query app at calcofi.io/db-query, or erddap.calcofi.io for filtered subsets.
Why these pages exist. Cloud storage serves objects, not directories — a bucket root returns an XML listing with no styling, and a URL ending in / has no listing at all. Each browsable folder here is a generated index.html, so the structure is navigable rather than guessable.