One oceanographic input file with all times to run ./cms

[Elisabet Cruz]

Hi CMS users!!

I am working with CROCO output data to run CMS. This model gives one NetCDF output file with all the times for the oceanographic data. I know that CMS need to have the data inside the file structured in a particular way and with one file for each time, but it will be easier if CMS will only work with one input file that have all the times to avoid the symbolic links to thousand of files.

Do you know if that is possible?

Have anyone tried to modified the Fortran codes to run cms reading the data from one only oceanographic input file?

I was also trying to modified getdata (mod_getdata.f90) to read my CROCO output data already transform from sigma to z-coordinates to generate one only file with all the times with the CMS structure, but this is not being so simple.

Thanks,

Elisabet

[bexeross]

Hi Elisabet, 

Sorry for a late reply. 

In case you don't know already, this forum is run/used mostly by other users. As we are not the original coders of CMS we are often not advanced enough in fortran to attempt the kind of modifications you are suggesting. (I may just be speaking for myself here, I don't know, but I personally am only confident to make minor tweaks, and am very clear on my own limitations within this field! Kudos to the original coders!)

I would agree that it is a pain to generate hundreds of symlinks and it would be good to get CMS working with a single file, and even better to get it to do the conversion from sigma to z format, but currently these are not capabilities of the software. 

So I would advise:

- adding a suggestion to the github issues page for the people who are actually working on the code (https://github.com/beatrixparis/connectivity-modeling-system/issues), although I imagine there is a reason that these changes aren't yet implemented (!).

- if you are more advanced in fortran than I, then do give it a go yourself, and if you succeed then please do share your modified class files both here and at the github so that others may use them too and that they may be integrated into the CMS codebase

- if you are getting stuck then I'm afraid the best thing to do is to follow the annoying instruction and make those symlinks (I assume that there could be a fast way to do this, that could be worth scripting a solution to). You will also need to do the conversion from sigma to z level yourself. I got help from an oceanographer (I'm a biologist) and made a script in matlab to do this conversion for me. Indeed you could use this script to generate one file per timestep if you wanted, but I suppose a plethora of symlinks is more space efficient.

You may have already figured all this out, and I'm sorry I can't be more help. But I can at least share in your frustrations and commend you for trying to modify the code to overcome these issues (even if you have to give up).

Good luck,

Bex