GeoZarr

STAC and Zarr Workshop Update (April 2025)

Goals

  • šŸ“ Provide a status update on the GeoZarr specification
  • 🧭 Provide entry-points for contributing to GeoZarr development

What is GeoZarr?

A geospatial extension for the Zarr specification, formalizing metadata expectations.

GeoZarr repository

Some history

GeoZarr commit activity

Recent activity

Christophe Noƫl proposes structuring GeoZarr around the Unified Abstract Data Model

Unified data model presentation

(Very) recent activity

Christophe Noƫl structures the spec based on the Unified Data Model and Conformance Classes

Unified data model proposal

Background

Unidata’s Common Data Model is an abstract data model that merges the NetCDF, OPenDAP, and HDF5 data models to create a common API for many types of scientific data.

Common data model

Unified Abstract Data Model

A Unified Abstract Data Model would encapsulate traditional models from the modelling (NetCDF, CF) plus models from the earth observation (GeoTIFF, GDAL) communities

Unified data model proposal

N-D functionality with E-O optimizations

Unified data model proposal

Unified Abstract Data Model Implementation

The VirtualiZarr library could be an implementation of the Unified Abstract Data Model

Virtualizarr diagram

Upcoming activity

Discussion about conformance classes April 17

Calendar event

Upcoming activity

Monthly GeoZarr meeting on May 7

Calendar event

Upcoming activity

Cloud Native Geospatial conference including a STAC + Zarr presentation by Julia Signell

Calendar event

Upcoming activity

EGU General Assembly including a EOPF Zarr discussion at the Pangeo session

Calendar event

Shameless plug

Pangeo showcase in 3 hours will include a presentation about new Zarr optimizations!

Calendar event

Shameless plug

Pangeo showcase in 3 hours will include a presentation about new Zarr optimizations!

Performance difference

Thanks! šŸš€

GitHub source