Skip to content

Store KGOs for CMIP7 ancillaries #387

@penguian

Description

@penguian

See issue #146 and ACCESS-NRI/CMIP7-Input#120

Ideally the CMIP7 ancillary suite for ACCESS-ESM1.6 should be able to check against Known Good Outputs (KGOs) to ensure that any deliberate changes to the suite do not cause unintended changes to the suite output.

Suite outputs are of two types: (1) files in the Gadi file system, and (2) branches of the https://github.com/ACCESS-NRI/access-esm1.6-configs repository.

For type (2) outputs, each new branch is created from an existing branch which acts as the KGO, and git diff can be used for comparison.
For type (1) outputs, I propose that there be a kgo/CMIP7/esm1p6_ancil directory in the gadi:/g/data/vk83 file system that stores copies of approved generated ancillary file directories. An example of such a directory is:

[pcl851@gadi-login-01 ~]$ tree /g/data/tm70/pcl851/CMIP7/esm1p6_ancil/2026.01.09.50
/g/data/tm70/pcl851/CMIP7/esm1p6_ancil/2026.01.09.50
└── modern
    ├── historical
    │   ├── atmosphere
    │   │   ├── aerosol
    │   │   │   └── global.N96
    │   │   │       └── 2026.01.09
    │   │   │           ├── BC_1849_2023_cmip7.anc
    │   │   │           ├── Bio_1849_2023_cmip7.anc
    │   │   │           ├── OCFF_1849_2023_cmip7.anc
    │   │   │           └── scycl_1849_2023_cmip7.anc
    │   │   ├── forcing
    │   │   │   └── resolution_independent
    │   │   │       └── 2026.01.09
    │   │   │           ├── TSI_CMIP7_ESM
    │   │   │           └── volcts_cmip7.dat
    │   │   └── land
    │   │       └── biogeochemistry
    │   │           └── global.N96
    │   │               └── 2026.01.09
    │   │                   └── Ndep_1849_2023_cmip7.anc
    │   └── forcing
    │       └── global.N96
    │           └── 2026.01.09
    │               └── ozone_1849_2023_cmip7.anc
    ├── historical-emissions
    │   └── atmosphere
    │       └── forcing
    │           └── global.N96
    │               └── 2026.01.09
    │                   └── CO2_fluxes_1849_2023_cmip7.anc
    └── pre-industrial
        ├── atmosphere
        │   ├── aerosol
        │   │   └── global.N96
        │   │       └── 2026.01.09
        │   │           ├── BC_1850_cmip7.anc
        │   │           ├── Bio_1850_cmip7.anc
        │   │           ├── OCFF_1850_cmip7.anc
        │   │           └── scycl_1850_cmip7.anc
        │   └── land
        │       └── biogeochemistry
        │           └── global.N96
        │               └── 2026.01.09
        │                   └── Ndep_1850_cmip7.anc
        └── forcing
            └── global.N96
                └── 2026.01.09
                    └── ozone_1850_cmip7.anc

What would be the process for creating and using these KGOs? How would checksums and manifests help?

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions