Recommendations for Safe Lossy Compression of weather and climate data

What lossy compression is safe when using lossy compression on weather and climate data?

This repository contains community-provided recommendations that are automatically compiled into one machine-readable recommendations.yaml file. These recommendations can then be used by external tools to automatically recommend appropriate safety requirements for compressing various weather and climate data.

This repository also provides the compression-recommendations Python package for loading these recommendations and inspecting them in the strongly typed Recommendations data structure.

Furthermore, we provide the following integrations:

• compression-requirement-checks checks whether safety requirements are upheld by a lossy-decompressed reconstruction of the original data.
• compression-requirement-safeguards translates safety requirements into compression safeguards that can be wrapped around any compressor to guarantee that the requirements are fulfilled.

Contributing: What should be recommended?

That is up to you, the community, to decide!

• You can add comments to any recommendation to elaborate or discuss different requirements.
• You can propose edits to any recommendation, which we accept once a new consensus has been reached.
• You can create new recommendations for new variables or to specialise existing requirements for specific cases.
• You can merge multiple recommendations to generalise them.
• You can propose new filters to select which recommendations should apply, or new kinds of requirements.

Glossary

Compression : Reducing the number of bits needed to store some data.

Lossy Compression : Compression that may only produce an approximation of the original data during decompression.

Filter : Criteria, or a composition of multiple criteria, that select which requirements should apply.

Safety Requirement : A property, e.g. an error bound, or a composition of multiple properties, that the lossy-compressed data must fulfil with respect to the original uncompressed data.

Citation

Please cite this work as follows:

Tyree, J., Faghih-Naini, S., Klöwer, M., Dueben, P., and Järvinen, H. (2026). Recommendations for Safe Lossy Compression of weather and climate data. Zenodo. Available from: doi:10.5281/zenodo.22761209.

Please also refer to the CITATION.cff file and refer to https://citation-file-format.github.io to extract the citation in a format of your choice.

Funding

The recommendations their accompanying compression-recommendations, compression-requirement-checks, and compression-requirement-safeguards packages have been developed as part of ESiWACE3, the third phase of the Centre of Excellence in Simulation of Weather and Climate in Europe.

Funded by the European Union. This work has received funding from the European High Performance Computing Joint Undertaking (JU) under grant agreement No 101093054.