Climate REF (Rapid Evaluation Framework)

Looking for results? Evaluations of CMIP7 data are published on the REF dashboard. This documentation is for people running the REF or developing diagnostics for it.

The Rapid Evaluation Framework (REF) is a set of Python packages that provide the ability to manage the execution of calculations against climate datasets. The aim is to be able to evaluate climate data against a set of reference data in near-real time as datasets are published, and to update any produced data and figures as new datasets become available. This is somewhat analogous to a CI/CD pipeline for climate data.

Installation

pip install climate-ref

If you want to use the diagnostic providers for the Assessment Fast Track, you can install them with:

pip install "climate-ref[aft-providers]"

Quick Start

# Ingest some observation datasets
ref datasets fetch-data --registry obs4ref --output-directory datasets/obs4ref
ref datasets fetch-data --registry sample-data --output-directory datasets/sample-data

# Run metrics against your climate data
ref solve

Features

• Real-time evaluation of climate datasets
• Support for multiple metrics providers (PMP, ILAMB, ESMValTool)
• Distributed processing capabilities
• Extensible architecture for adding new metrics providers
• Command-line interface for easy interaction

Documentation

For detailed documentation, please visit https://climate-ref.readthedocs.io/

Contributing

REF is a community project, and we welcome contributions from anyone. Please see our Contributing Guide for more information.

License

This project is licensed under the Apache License 2.0 - see the LICENSE file for details.