Clean up run_tower.py

[TeaganKing]

In a discussion with @danicalombardozzi , Sam, and Gordon, we talked about a few items related to run_neon.py that would ideally be changed for clarification purposes. Rachel has run into a few confusing pieces of information while working with Danica, so we might want to keep her in the loop and/or tag her if she makes a GitHub account.

Here is our ideal tower refactoring wish list (for after PLUMBER2 comes in):

• Clarify the use of base case. This is confusing for users because we use the same base case in multiple places.
• Start date is confusingly set to year 18 in env_run.xml based on NEON usually starting in 2018 and ad/postad starting on 0018 and 1018 respectively. This doesn't necessarily cause issues, but is confusing.
• AD spin up is set up with --setup_only but then requires ./case.setup which could be misleading
• There is a duplication of code with user_nl_clm and changes in the first bit of the namelist are overwritten by the second instance of the namelist
• Use RUN_REFTOD instead of START_TOD in PLUMBER2 usermods & update buildnml tod accordingly

[danicalombardozzi]

Here is another issue to explore. Setting up a transient simulation (for MOAB, haven't tested other sites) using the --experiment option set incorrect values for these variables:
entry id="RUN_STARTDATE" value="2000-01-01"
entry id="STOP_N" value="5"

[TeaganKing]

Thanks Danica.

I'll also add the BGC/SP case flag as an additional feature to address on this issue (see discussion in #2406).

[ekluzek]

These look like things that would be good to handle.