Open research data, built to be moved in bulk.
Northmere Data Commons keeps 2,184 collections of genomics, climate reanalysis and satellite imagery, 11.6 PB in total, and serves them over HTTPS, aria2 manifests and Globus-style transfer endpoints. Every collection has a persistent identifier, a licence and a checksum list.
Twelve collections, from 3 TB to 80 TB
Imagery and climate archives dominate the top of the list; the long tail of smaller genomics releases is what most visitors download first.
Three kinds of data, one way to fetch them
Each domain keeps its own curators and release calendar. Files are plain, documented formats with checksums, so a transfer can resume or be verified without our tools.
Genomics
Raw reads, long-read assemblies, pangenome graphs and single-cell atlases. Large cohorts are published as numbered releases and never edited in place; corrections are new versions with a changelog.
- Collections
- 612
- Volume
- 3.4 PB
- Typical format
- CRAM, FASTQ.gz, Parquet
Climate and ocean
Atmospheric and ocean reanalyses, regional downscaling ensembles and sea-ice records, chunked as Zarr or NetCDF-4 so a single variable and time range can be read without fetching the rest.
- Collections
- 487
- Volume
- 4.1 PB
- Typical format
- Zarr v3, NetCDF-4
Satellite and aerial imagery
Optical and radar tiles on one fixed grid, stored as cloud-optimised GeoTIFF with a STAC catalogue. Ranged reads make it practical to pull one window out of a 480 MB tile.
- Collections
- 1,085
- Volume
- 4.1 PB
- Typical format
- COG, STAC, GeoParquet
Fetch a collection from a terminal
Registered users get a signed manifest for any collection: one line per file with a size, a SHA-256 and a mirror list. Feed it to aria2 and the transfer resumes after any interruption.
Anonymous visitors can still download over plain HTTPS at 400 Mbit/s with four connections. Researcher accounts are free and raise the limit to 2 Gbit/s and 12 TB each month.
# signed manifest for the soil metagenome survey, 1,862,400 files
curl -fsSL -H "Authorization: Bearer $NM_TOKEN" \
https://se2.enotcloud.com/files/gen-0388/MANIFEST.aria2 -o gen-0388.aria2
aria2c -i gen-0388.aria2 -x 16 -s 16 -j 8 \
--check-integrity=true -d ./gen-0388
# 63.4 TB at 5 Gbit/s sustained is about 28 hoursWhat the commons served in the last 90 days
1.92 PB to 14,310 institutions in 96 countries during October. Bandwidth is donated by member networks and sized for release days.
Persistent identifiers
Every collection version has an identifier of the form 10.88213/nmdc.clm.1120.v5.2 that resolves to its landing page for as long as the consortium exists, and to a tombstone page afterwards.
Four serving regions
Files are mirrored across nm-eu-north, nm-eu-central, nm-us-central and nm-ap-south. Requests are routed to the nearest healthy region and fall back automatically. See the status page.
Funded by members
The service is paid for by 38 member institutions and four research funders, not by advertising or data sales. The funding page lists every contribution above 1 million euro.
News from the release desk
Release 2026.3: SAT-2031 version 6.0
Eight years of tiles, 79.8 TB, new processing baseline.
NMR-5 reaches 2025
The atmospheric reanalysis now covers 1950 to 2025.
840 new soil sites in GEN-0388
The 2025 field season adds 9.1 TB of reads.
Join the consortium
Universities, institutes and national data services can join as associate or full members. Members share an egress allowance, sit on the technical committee and help decide which collections the commons takes in next.
- Associate
- 21,000 EUR per year, 300 TB per month pooled
- Full member
- 72,000 EUR per year, 2 PB per month pooled
- Data node
- Host a regional mirror and receive full membership