Namelist improvements

[SeanBryan51]

High level issue for tracking namelist file improvements:

• cleanup this ridiculous l_casacnp switch #166
• Add deprecated namelist options in the deprecated feature list #137
• USER guide namelist descriptions #143
• Setting icycle > 10 #659
• Rename namelist option output%averaging #661
• @JhanSrbinovsky proposed some namelist changes in Nml draft3 #588

[JhanSrbinovsky]

@SeanBryan51 - I never seem to have the time to get back to this - that first one #166 - this has been implemented in ESM1.6, maybe AM3 as well, I can't recall but in both of those cases the IO is done at the front end - actually maybe the other way around:
Here's the file 1st SUBR in the module

[JhanSrbinovsky]

Coupled - the namelists are in a rose configuration file, which then writes them out as .nmls for the model to read. It makes particular sense there, but in general I think it makes sense for each science module to have its own namelist and then each other technical thing. e.g. there would be a canopy.nml, radiation.nml an output.nml etc. The coupled model doesn't need output.nml at all. CM3 doesn't need casa.nml etc. cable.nml has been a junk drawer for 20 years. cable_user% was presented as a template for developers on how to create switches that could be turned on in a namelist and picked up any where via a module. I didn't anticipate developers would just start adding to cable_user TYPE directly.

[JhanSrbinovsky]

@SeanBryan51 - In my first comment the linked SUBR/module - if you can do something self-contained like that then I can pick it up easily in the coupled model as well and we don't have to repeat the process

[SeanBryan51]

Thanks Jhan! I'm not too familiar with the state of CABLE namelists outside the offline context so this is all very useful