Sample data for use with ESMValTool

This repository will contain samples of real data for use with ESMValTool for automated testing and possibly for demonstration purposes. The goal is to keep the repository size small (~ 100 MB), so it can be easily downloaded. We strive to include data that adheres well to the CMIP6 standard and can be loaded using simple iris commands.

At present, the repository contains 44 monthly, and 35 daily timeseries datasets. The specifications can be found in datasets.yml.

The data are stored using the data reference syntax structure from DKRZ.

Usage

The functionality of this repository is kept simple.

After installing iris, this package can be installed with the command:

pip install ESMValTool_sample_data

Data can be loaded using the following command for daily or monthly timeseries data.

import esmvaltool_sample_data

# load monthly air temperature timeseries data
ts_amon_cubes = load_timeseries_cubes(mip_table='Amon')

# load daily air temperature timeseries data
ts_day_cubes = load_timeseries_cubes(mip_table='day')

config-user.yml

If you want to use this dataset in your ESMValTool projects, add the following lines to your config-user.yml:

rootpath:
  CMIP6:
    - <path_to_repository>/esmvaltool_sample_data/data/timeseries/CMIP6

drs:
  CMIP6: DKRZ

If you have installed esmvaltool_sample_data, you can find the rootpath settings using python -c 'import esmvaltool_sample_data, yaml; print(yaml.dump(esmvaltool_sample_data.get_rootpaths()))'

License

This work is licensed under Apache 2.0 (code) and CC-BY-SA 4.0 (data). All data files in the directory (esmvaltool_sample_data/data/) are derived from CMIP6. The licensing agreements governing CMIP6 data depend on the model but generally conform to CC-BY-SA 4.0, see the CMIP6 Terms of Use for more detailed information. The terms of the Apache 2.0 license are available in the LICENSE file, and the terms of the CC-BY-SA 4.0 license in the esmvaltool_sample_data/data/LICENSE file.

How to contribute

Suggestions/improvements/edits are most welcome. Please read the contribution guidelines before creating an issue or a pull request.