Update CMEPS/MOSART/CISM/RTM tags

[ekluzek]

Update MOSART, CMEPS and CISM so CISM runoff goes to ROF rather than CTSM
Update RTM with fix needed for Paleo LGM work

[slevis-lmwg]

From meeting with @ekluzek

• Set separate_glc2ocn_fluxes = .false. and rerun the test-suite: expect bfb to baseline:
  FAIL ./run_sys_tests -s mosart -c mosart1.1.01_ctsm5.2.007 --skip-generate
  Same failure as with separate_glc2ocn_fluxes = .true.
  Update: I checked off this TODO because it is not applicable at this point.
• Update the submodules tags to the branches: cism runoff will be now routed to ocn via mosart MOSART#94 and the 3 listed here). At this time do not need the ccs_config and rtm updates. The former (ccs_config) is ready so bring in regardless. The latter (rtm) is pending so bring in when it's time to merge to master. Update: At this point the ccs_config and the rtm PRs are ready, so I'm bringing them in.
• Update .gitmodules temporarily: url (my own) and tag (try the branch name or may have to point to the hash on my fork).
• Open PR for my ctsm branch from master.
• @ekluzek will review cism runoff will be now routed to ocn via mosart MOSART#94 to discuss possible code changes and likely need for cdeps
• Make a branch tag for this branch (same process as we followed for the ctsm5.2 branch, see naming convention, branch_tags/cismtorof.n01_ctsm5.2.007)
• Add 2 new tests to the mosart test-suite:
  Start with ERS_Ld5.f10_f10_mg37.I2000Clm60BgcCru.derecho_gnu.mosart-default both using cism compsets, one setting separate_glc2ocn_fluxes = .false. in its tesmod directory and the other using the default (so no testmod needed).
  Update: This namelist option has been removed, so I will now add only 1 new test to include cism in the compset.
• Run the mosart test-suite and generate baseline mosart1.1.02_ctsm5.2.007:
  derecho and izumi OK: NLCOMP and "field lists differ"
  derecho sanity check OK: ./run_sys_tests -s rtm -c rtm1_0_80_ctsm5.2.007 --skip-generate NLCOMP only
• Run aux_clm to make sure running with cism works and generate baseline cismtorof.n01_ctsm5.2.007
  derecho and izumi OK: NLCOMP and "field lists differ" BUT ALSO DIFF from baseline for expected vars (e.g. TWS, *volr, *volrmch, *rofl)
• Make the mosart tag and the rtm tag:
  cism runoff will be now routed to ocn via mosart MOSART#94 mosart1.1.02
  fix area scaling to correct RTM river flux to MOM6 RTM#51 rtm1_0_80
• @jedwards4b makes tags:
  Implementation of glc->ocn coupling via going through mosart  CMEPS#463 cmeps0.14.67
  removal of mapping files that are no longer needed for cism->mosart ESMCI/ccs_config_cesm#174 ccs_config_cesm0.0.110
• @Katetc makes the cism tag.
• Update to the new tags
• Run ./run_sys_tests -s aux_clm -c cismtorof.n01_ctsm5.2.007 --skip-generate
  Run ./run_sys_tests -s ctsm_sci -c ctsm_sci-ctsm5.2.007 --skip-generate
• Results and waiting for feedback in Update submodule tags to pass runoff from CISM to ROF as per issue #2590 #2605 (comment)
• Repeat testing with the latest master tag, merge the PR, and tag as normal.

[slevis-lmwg]

@ekluzek our first TODO with separate_glc2ocn_fluxes = .false. failed the same way as separate_glc2ocn_fluxes = .true. so this may need to be part of your code review and conversation with Mariana.

[slevis-lmwg]

In the above TODO list, I just checked off "Make a branch tag" by doing:

git tag -a branch_tags/cismtorof.n01_ctsm5.2.007
git push escomp branch_tags/cismtorof.n01_ctsm5.2.007 

[slevis-lmwg]

Mosart and rtm tags made.

@jedwards4b and @Katetc
We now need tags for

• Implementation of glc->ocn coupling via going through mosart  CMEPS#463
• change glc -> ocn export fields to be glc->rof CISM-wrapper#100
• removal of mapping files that are no longer needed for cism->mosart ESMCI/ccs_config_cesm#174