Icechunk ERA5

Data coverage

An analysis-ready, cloud-optimized (ARCO) edition of ERA5, the fifth-generation global reanalysis from the Copernicus Climate Change Service (C3S) at ECMWF. Provided as a single Icechunk + Zarr v3 store on AWS S3, this version enables efficient, scalable access to global reanalysis data for weather, climate, and ML/AI applications. Managed by Earthmover. ERA5 provides hourly estimates of atmospheric, land-surface, and ocean variables from 1940 to 2025 on a ~31 km (0.25°) global latitude-longitude grid. This Earthmover edition delivers 35 single-level (surface) variables, 8 pressure-level variables on 13 isobaric levels (50–1000 hPa), and a curated 500 hPa geopotential slice, all with CF-1.7 compliant metadata. The store is organized into three top-level groups (single, pressure, and 500hPa). Each group is provided with two complementary chunking strategies that hold identical data: a "spatial" layout (one full global map per hour) optimized for map-style and regional queries, and a "temporal" layout (long time series in small spatial tiles) optimized for time-series extraction, location-based analysis, and ML training workflows. Arrays are stored as Float32 with PCodec compression for high compression ratios. Only lossless compression has been applied; data values are bit-for-bit identical to raw ECMWF NetCDFs. The data is provided in the Icechunk format. Icechunk is an open source tensor storage engine that augments Zarr with transactional updates, version history, and an optimized, Rust-based I/O path for maximum performance in the cloud. An interactive browser for the data is available via the Earthmover Data Marketplace: https://app.earthmover.io/marketplace/6a19bcfe9aa6e97720a2fad2?tab=dataset Earthmover maintains two complementary ERA5 datasets: - This one, updated at quarterly frequency, available free under an open license. - A low-latency version with the same variables, but updated daily with the latest ERA5T near-real-time data. This is available at a competitive price and comes with a strong SLA. See ERA5 (Daily Updates) on the Earthmover Data Marketplace. If your organization is using ERA5 in a serious operational context, please consider subscribing to the low-latency version of this dataset. Paid subscribers for the low-latency dataset are what make it sustainable for us to provide this high-quality free, open dataset. Contact [email protected] to learn more. The ERA5 dataset is public and can be opened directly from S3 with the Icechunk and Xarray libraries: python import icechunk import xarray as xr storage = icechunk.s3storage( bucket="earthmover-icechunk-era5", prefix="icechunkV2", region="us-east-1", anonymous=True, ) repo = icechunk.Repository.open(storage) session = repo.readonlysession("main") ds = xr.openzarr(session.store, group="single/spatial", consolidated=False, chunks=None) Alternatively, you can use the Arraylake client for a more concise syntax. python from arraylake import Client import xarray as xr client = Client() client.login() repo = client.getrepo("earthmover-public/era5") session = repo.readonlysession("main") ds = xr.openzarr(session.store, group="single/spatial", consolidated=False, chunks=None) The dataset is assembled from the NSF NCAR Curated ERA5 archive and the Copernicus Climate Data Store. It is widely used for ML/AI training, climate and weather research, extreme-event analysis, and applications such as renewable energy siting.

Access & permitted use

Review the publisher’s documentation for availability, eligibility and permitted uses, including commercial analysis, model training and redistribution.

Creative Commons Attribution 4.0 International (CC-BY 4.0) - https://creativecommons.org/licenses/by/4.0/

Need help sourcing data?

Tell us your required fields, population, geography and time period. We can review suitable sources and licensing options.