Source data and attribution

Every fixture is a lossless re-encoding of one array from the EarthCompress benchmark. manifest.json records the exact corpus, sample ID, and original array checksum.

Sentinel-2 L1C

The samples originate from COPERNICUS/S2_HARMONIZED in Google Earth Engine.

Required attribution: Contains modified Copernicus Sentinel data 2023.

Sentinel-1 GRD

The sample originates from COPERNICUS/S1_GRD in Google Earth Engine.

Required attribution: Contains modified Copernicus Sentinel data 2023.

ERA5 2 m temperature

The sample originates from the public ARCO-ERA5 Zarr archive on Google Cloud.

Required notice: Contains modified Copernicus Climate Change Service information 2023. Neither the European Commission nor ECMWF is responsible for any use of this data.

AlphaEarth Foundations Satellite Embedding

The sample originates from GOOGLE/SATELLITE_EMBEDDING/V1/ANNUAL, dataset version 1.1 and model version 2.1. The dataset is licensed under CC BY 4.0.

Required attribution: The AlphaEarth Foundations Satellite Embedding dataset is produced by Google and Google DeepMind.

EMIT L2A reflectance

The sample originates from NASA/EMIT/L2A/RFL in Google Earth Engine. NASA EMIT data may be used and redistributed without restriction. Cite the EMIT L2A Reflectance product when publishing derived work.

ESA WorldCover 2021 v200

The sample originates from the public ESA WorldCover 10 m Cloud Optimized GeoTIFFs.

Attribution: © ESA WorldCover project 2021 / Contains modified Copernicus Sentinel data (2021) processed by the ESA WorldCover consortium. The dataset is licensed under CC BY 4.0.