This special issue publishes model evaluation papers emerging from the Coupled Model Intercomparison Project Phase 6 (CMIP6) and other model intercomparison projects (MIPs). CMIP evaluation papers present particular challenges both for authors and for GMD’s policy on code and data availability.
This special issue interprets the code and data availability rules to make them applicable to CMIP evaluation papers and to provide clarity to authors as to what they need to provide in their manuscripts.

The guidance here must be read in conjunction with the GMD rules on manuscript types and code and data availability.
https://www.geoscientific-model-development.net/about/manuscript_types.html https://www.geoscientific-model-development.net/about/code_and_data_policy.html
The rules there must be followed, subject to the specific provisions outlined below.

Manuscript scope
The manuscript must meet the scope rules for a type of GMD manuscript. This will usually be “model evaluation” but may be another GMD-recognised type. In particular, this means that the subject of the manuscript must be the performance, behaviour, and/or design of the models. Manuscripts whose subject is the behaviour of some aspect of the Earth system or which draw conclusions about that system will be rejected as out of scope. Instead, they should be submitted to appropriate application area journals.

Publication of the MIP protocol
The experimental protocol for the MIP being evaluated must have been published in a peer-reviewed forum in advance of the MIP being conducted and must be cited in the manuscript. For CMIP6, this will have been in the GMD special issue “Coupled Model Intercomparison Project Phase 6 (CMIP6) Experimental Design and Organization”.
Where the evaluation uses model results which were produced in circumstances which deviate from the published protocol, this deviation must be documented either in the evaluation manuscript or in another paper cited in the manuscript.

Model code and data requirements
MIP evaluation papers will frequently span simulations conducted with many models, most or all of which are outside the control of the authors conducting the evaluation. The interpretation of the code and data availability rules presented here is designed to ensure the transparency and reproducibility of the evaluation itself.
1. All of the model outputs and observational data used in the evaluation must be publicly and persistently archived in accordance with GMD policy. For CMIP6, this will usually be on Earth System Grid Federation (ESGF). The manuscript must identify the data used with sufficient precision that a reader would be able to reproduce the analysis. GMD policy explicitly prohibits embargoes. If the data are embargoed, then manuscript submission must be delayed until the embargo expires.
2. The exact version of the analysis code used to conduct the evaluation must be publicly and persistently archived, subject only to the exceptions beyond the authors’ control which are detailed in the GMD code and data availability policy.
3. In the case of evaluations of multiple models participating in the same MIP, the usual requirement that there must be a published model description paper about the models being evaluated is waived. Where the evaluation is of a single model’s performance in a MIP, the usual requirement that there must be a previous model description paper applies.