Use Cases
Faster analysis workflows, from air quality to oceanography to fields well beyond Earth science.
TEMPO Air Quality Analysis
The TEMPO instrument measures air pollutants over North America during daylight hours, producing ~10 Level 2 netCDF files and one Level 3 file per hour. A year's worth of Level 3 data totals ~5,000 files (~2.5 TB).
With DMR++ parsing, VirtualiZarr, and earthaccess, researchers on the Openscapes JupyterHub opened the entire year as granule-level chunk manifests and merged them into one datacube.
Full-year access in ~10 minutes instead of about a day
In the wild
Production pipelines, living stores, and integrations built on virtual Zarr, from NASA DAACs to the ESGF to single-cell biology.
GOES-16 satellite imagery
Seven years of GOES-16 imagery, 380,000 netCDF4 files on a public bucket, opened as one cloud-optimized virtual Zarr store.
Learn moreGPM IMERG precipitation
A pipeline building a virtual Icechunk store over the GPM IMERG half-hourly precipitation record, nearly half a million granules.
Learn moreMUR sea surface temperature
Recurring updates append newly published MUR SST granules to an ongoing Icechunk repository.
Learn moreGEOS-CF composition forecasts
Virtual Icechunk stores for NASA's GEOS-CF dataset, hosted publicly on S3.
Learn moreNEX-GDDP-CMIP6 projections
NASA's NEX-GDDP-CMIP6 downscaled climate projections exposed as a single virtual datatree.
Learn moreCMIP7 at the ESGF
An ESGF working group enabling access to CMIP7 climate model output via virtual Zarr.
Learn moreEUMETSAT firecube
A plugin-based batch ingestion CLI turning Earth Observation products into analysis-ready datacubes.
Learn morePO.DAAC operations
PO.DAAC operationalizing its cloud-optimized format generation, with virtual Zarr as DAAC infrastructure.
Learn moreSingle-cell biology
A collection of h5ad/AnnData files referenced as one virtual sparse matrix in Icechunk, with no copying.
Learn moreBuilt on virtual data
The consumer side: analysis, benchmarking, serving, and cataloging tools that read virtual stores rather than publish them.
ExtremeWeatherBench
Brightband's benchmark of machine learning and numerical weather prediction models, focused on extreme events.
Learn moretitiler-multidim
A TiTiler application serving map tiles directly from netCDF and Zarr datasets.
Learn moreintake-virtual-icechunk
ACCESS-NRI's intake plugin for building and reading Icechunk stores from existing ESM datastores via VirtualiZarr.
Learn moreTEEHR
RTI International's tools for exploratory evaluation in hydrologic research.
Learn morepm25ml
CREA's two-stage machine learning model producing daily 10 km PM2.5 estimates across India.
Learn morearco-ocean
A dataset of physical ocean, wave, and sea-ice properties with hydrological and atmospheric forcing, optimized for machine learning.
Learn moreVirtual datasets, productized
The Earthmover Data Marketplace serves analysis-ready datacubes in the open Icechunk format, including virtual datasets built over existing netCDF, HDF5, GRIB, and TIFF archives.
Try it yourself
Start with VirtualiZarr and earthaccess to bring cloud-native workflows to your research.