Zarr at Scale

Appearance

Zarr at ScaleWhat's new in Zarr

Working with multi-terabyte Earth-science datasets is now routine.

View the poster
Download as PDF
Read the abstract
View on GitHub

Sharding

Many small chunks, few storage objects. Random access of small chunks with the storage footprint of large ones.

Virtualization

Read NetCDF4, HDF5, GRIB, and GeoTIFF archives as one Zarr dataset, no conversion or copies.

Variable chunk grids

One Zarr array, many chunk sizes. Just landed in zarr-python.

In-browser rendering

Stream Zarr chunks straight to the GPU with Zarrita.js and deck.gl-zarr. No server, no pre-rendered tiles.

GeoZarr

Open conventions for geospatial Zarr. CRS, spatial transforms, multiscale pyramids, on the OGC standards track.

EGU 2026 poster ​

This site is the companion to the EGU 2026 poster Zarr at scale: virtualization, sharding, and performance optimizations for Earth science data (EGU26-15196), presented in session ESSI2.2 on Monday 4 May 2026.

See also ​

Citation ​

Jones, M., Hamman, J., Bennett, D., Barron, K., and Magin, J.: Zarr at scale: virtualization, sharding, and performance optimizations for Earth science data, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-15196, https://doi.org/10.5194/egusphere-egu26-15196, 2026.

Acknowledgments ​

Max Jones' Zarr contributions were supported by the Data Systems Evolution (DSE) team at NASA Marshall Space Flight Center's Office of Data Science and Informatics (ODSI). DSE enables scientific exploration and discovery through innovative data visualization techniques and analysis capabilities that lower the barrier to entry for cloud-hosted data.

EGU 2026 · ESSI2.2 · EGU26-15196

CC BY 4.0