SMYLE use-case setup

[ekluzek]

Bring in a user-mod directory on the release-clm5.0 branch that implements the following for CLM:

https://docs.google.com/document/d/1rV3fWCIt-D1fAbzQ6uaQeZyLagzzKnopofr88eXegzY

[ekluzek]

I'll also work with someone on CAM for this

ESCOMP/CAM#256

[ekluzek]

I have a branch tag for this now: branch_tags/smyle.n01_release-clm5.0.34

[ekluzek]

OK, we need a user mod for 1850 and a separate one for Historical. And for 1850 we need a TRENDY surface dataset for that as TRENDY used 1700. I've done that and created a new tag: branch_tags/smyle.n02_release-clm5.0.34

[ekluzek]

Working with @nanr we found we needed to separate out the versions for stand-alone and coupled (otherwise the coupled one fails when trying to set DATM settings and especially trying to copy a datm namelist changes in SourceMods). I also changed it so that the user-mod is set when a %SMYLE setting is used for the compset name. This version is branch_tags/smyle.n03_release-clm5.0.34

[ekluzek]

We decided we should further work on this and make the smyle non-history changes into a use-case so it's more embedded into the namelist build. This means you don't have to append further changes onto the user_nl_clm, but can do a complete replacement of the user_nl_clm without losing anything.

@nanr also found that there was a typo in the filenames for the TRENDY files.

[ekluzek]

@jedwards4b has a PR that changes the 1850 user-mod directory into a use-case for SMYLE. #1614. I think the other two user-mod directories need to be handled similarly.

[ekluzek]

The PR #1328 implemented this using user-mod directories. That approach was complex and the use-cases in #1614 are a better way to do it. At this point I think everything in the first is in #1614, although that could be checked. But, we need to make sure these are correct for what needs to be done for SMYLE.