Supported model data

[roberts-luke]

Its my understanding that model data needs to be CMORized for use with ESMValTool (https://docs.esmvaltool.org/en/latest/input.html#using-a-cmorizer-script) , and at the moment CMORization is available for observational data, rather than model data.

Is there any method to input local model data, which has not been used for CMIPs, into a format where ESMValTool can read and use it.

Thanks

[zklaus]

We have some support for native model data, more on that below.

Let me start by saying that by cmorized data, we generally mean data that follow the standards established for CMIP and used also in projects like CORDEX. That means first of all valid netCDF files that follow the CF Conventions and that also additionally follows the metadata requirements laid out in the CMIP6 Participation Guidance for Modelers, i.e. uses the correct variable names and coordinate encoding, etc.
If your data is not yet part of the CMIP6 Data Request, for example, because you have a new variable, perhaps a new chemical species or something similar, you may be able to apply your institutes normal cmorization procedure to produce cmorized files that can be read by ESMValTool with the help of a custom table. We'll be happy to discuss this prospect and help out with it for any given variable!

Coming back to the native model output, if all of this cmorization talk sounds a bit scary and too much, fret not. We have some support for directly reading native model data, particularly for netCDF, grib, and PP files. This is realized via the extra facets mechanism.

What use case do you have in mind? Do you have something like a traditional GCM/ESM with some extra variables? Or are you thinking about an in-situ soil model? What kind of variables are you interested in?

If you give us a bit of an idea about the concrete problem, we will probably be able to help you figure something out.