ICOS Carbon Portal · zarr datasets
These ICOS environmental datasets are published as Zarr — a cloud-native format for chunked, compressed N-dimensional arrays that clients such as xarray open directly over HTTP, reading only the chunks they need. ICOS Carbon Portal serves them here for automated workflows and AI-assisted data discovery and harvesting: each dataset is self-describing (schema.org + Croissant metadata, an agent guide, and a REST query API). This proxy provides a unified interface for accessing these datasets and mints RO-crate based data passports on access to enable transparent, reproducible and FAIR data usage, while making sure at the same time that users are informed about attribution, data license, variable metadata, data quality and provenance.
Datasets (1)
Explore — demonstration apps
The datasets are also browsable through interactive demonstration apps:
- ICOS multi domain data browser
- ICOS ObsPack atmosphere viewer
- FLUXNET Shuttle viewer
- SOCAT ocean viewer
- NOAA GlobalView atmosphere viewer (login restricted — contact info@icos-cp.eu)
Learn — sample notebooks
Executed end-to-end against the live services (outputs included), one per access route. They live — with standalone example scripts, the service reference and a walk-through tutorial — in the public CodingHelpAndExamples repository:
- 01 · Portal data objects (icoscp_core) — the citable route: discovery, citations, binary column reads
- 02 · Atmosphere CO₂ quickstart (Zarr) — a tall-tower series in five cells, QC flags, session passport
- 03 · REST query API + passports — filtered extracts by bbox/height/time, Arrow bulk reads
- 04 · Ecosystem fluxes (FLUXNET) — aggregation groups, NEE/GPP/RECO reference slices, fraction QC
- 05 · Ocean SOCAT cruises — ship tracks with reduce=track, one cruise in full, WOCE QC
- 06 · OPeNDAP for large grids — multi-GB netCDF sliced server-side — and when not to use DAP
- 07 · Reproducibility — passports on every route, integrity hashes, the results checklist
- 08 · Across stations (combined views) — dense panels: pick stations by geography, lazy latitude transect
- 09 · Zarr under the hood — store anatomy with zarr-python, whole-store inventory, gridded cube
- 10 · North Sea blooms (advanced) — a full worked study: bloom detection, thermal vs biological pCO₂, air–sea flux with ERA-5 winds
- 11 · Site explorer (advanced) — one site end to end: every group, provenance, 4-D METEOSENS profiles
- 12 · Cross-domain extraction (advanced) — one bbox + year across atmosphere, ecosystem and ocean — one expression each
- 13 · Global SOCAT tour (advanced) — 44M-obs synthesis: cruises, ships, the fCO₂ globe, the 25-year trend
- 14 · ObsPack network explorer (advanced) — station index, coverage Gantt for every station × gas, network map
More: standalone example scripts · service reference · tutorial
Machine access: llms.txt · OpenAPI · send Accept: application/ld+json for the raw catalog.