Distinguish between ocean and wetland in fsurdat files

[slevis-lmwg]

Description of changes

Add pctocn to mksurfdata_esmf and write PCT_OCEAN to fsurdat files.

Specific notes

Contributors other than yourself, if any:
@billsacks @ekluzek @mvertens

CTSM Issues Fixed (include github issue #):
Fixes #1886

Are answers expected to change (and if so in what way)?
No

Any User Interface Changes (namelist or namelist defaults changes)?
fsurdat contains PCT_OCEAN now

Testing performed, if any:
Code worked to generate the following fsurdat file /glade/work/slevis/git/mksurfdata_toolchain/tools/mksurfdata_esmf/surfdata_1.9x2.5_hist_78pfts_CMIP6_1850_c230914.nc
I will also need to change clm code to successfully read the new fsurdat file.

[slevis-lmwg]

Run all python tests

• Unit tests PASS

• Sys tests PASS

• ./run_sys_tests -s clm_pymods --skip-compare --skip-generate FAIL
  LILACSMOKE FAIL expected due to pre-pctocn fsurdat file (i.e., does not include PCT_OCEAN; same reason all aux_clm tests will fail until we replace all fsurdat files)
  FSURDATMODIFYCTSM FAIL likely expected because not finding conda in this tag

• add new /testinput files to repo

• do not test aux_clm (will fail)

Generated new 78-pft (also 16-pft) 1-deg fsurdat/landuse files

surfdata_0.9x1.25_hist_78pfts_CMIP6_1850-2015_c231031.namelist
surfdata_0.9x1.25_hist_78pfts_CMIP6_1850-2015_c231031.nc
landuse.timeseries_0.9x1.25_hist_78_CMIP6_1850-2015_c231031.nc 

and old (i.e. without PCT_OCEAN)

surfdata_0.9x1.25_hist_78pfts_CMIP6_1850-2015_c231107.namelist
surfdata_0.9x1.25_hist_78pfts_CMIP6_1850-2015_c231107.nc
landuse.timeseries_0.9x1.25_hist_78_CMIP6_1850-2015_c231107.nc

Ran these cases with "cold start" initial conditions:
/glade/u/home/slevis/cases_ctsm5.2/i1850_f09_bgccrop
/glade/u/home/slevis/cases_ctsm5.2/i1850_f09_bgccrop_default
and got bfb, including after a restart.

[slevis-lmwg]

Currently seeing diffs (not in day 1 but in day 2 and later) in a single grid cell: ncview indices (i,j) = (20,68) and (lat, lon) = (38.8,50)

I have compared the two fsurdat files used (old, without PCT_OCEAN, listed first and new listed second) and found different mapunits in the same grid cell:
/glade/work/slevis/git/mksurfdata_toolchain/tools/mksurfdata_esmf/surfdata_1.9x2.5_SSP5-8.5_78pfts_CMIP6_1850-2100_c230227.nc
'/glade/work/slevis/git/mksurfdata_toolchain/tools/mksurfdata_esmf/surfdata_1.9x2.5_hist_78pfts_CMIP6_1850_c230914.nc'

I generated a new copy of the "old" fsurdat file. Now mapunits are the same as in the new fsurdat.
I generated a new copy of the "old" as was done before, i.e. for transient. The mapunits are still the same as in the new fsurdat.

I will attribute the mapunit difference in that one grid cell to a change in the ctsm5.2 branch between tags
alpha-ctsm5.2.mksrf.15_ctsm5.1.dev123 and alpha-ctsm5.2.mksrf.14_ctsm5.1.dev111-5-gb1204ca46
or some change in cheyenne between late Feb and now.

[slevis-lmwg]

Notes to self:
While working on izumi, I will try this test when looking for bfb answers:
SMS.f10_f10_mg37.I2000Clm51BgcCrop.izumi_gnu.clm-crop
First generate new fsurdats, one without ocean and one with ocean.

[slevis-lmwg]

Last commit not tested on izumi because ./gen_mksurfdata_build.sh fails on izumi.

Testing on cheyenne with existing cases:
/glade/u/home/slevis/cases_ctsm5.2/i1850_f19_bgccrop_default
/glade/u/home/slevis/cases_ctsm5.2/i1850_f19_bgccrop

[slevis-lmwg]

My latest modifications have not changed answers.
@swensosc would you like to look at my last two commits for confirmation that I did what you suggested?
If you have no further comments, this PR can be merged with the ctsm5.2.mksurfdata branch.

[slevis-lmwg]

@swensosc thanks for your email (@swensosc looked at my code and was ok with my mods).
@swensosc suggested an additional test, where we work with a raw dataset that includes inland wetlands and confirm that everything works as intended. I like the suggestion.

Update:
The corresponding raw dataset (mksrf_lanwat.050425.nc) includes non-zero wetlands, so I just need to set no_inlandwet = .false. in mksurfdata_esmf for Sean's test. I generated the fsurdat file:

/glade/work/slevis/git/mksurfdata_toolchain/tools/mksurfdata_esmf/surfdata_0.9x1.25_hist_78pfts_CMIP6_1850_c231031.namelist
/glade/work/slevis/git/mksurfdata_toolchain/tools/mksurfdata_esmf/surfdata_0.9x1.25_hist_78pfts_CMIP6_1850_c231031.nc

Simulation pointing to this fsurdat /glade/scratch/slevis/i1850_f09_bgccrop/run completed successfully.

[slevis-lmwg]

Updating to latest ctsm5.2.mksurfdata...
Conflict in python/ctsm/test/testinputs/modify_fsurdat_short.cfg suggests that I may need to remake this file
ctsm/test/testinputs/surfdata_5x5_amazon_hist_16pfts_CMIP6_2000_c231031.nc to a 78-pft file, but we'll see if tests pass.

[slevis-lmwg]

PASS: make all
FAIL: clm_pymods same expected fails as before (see earlier post)