Updates to add new Catchment-CN version: Catchment-CN5.1

[gmao-jkolassa]

This PR introduces the new version Catchment-CN5.1 land model, based on the Community Terrestrial System Model version 5.1 (CTSM5.1). Most of the files related to Catchment-CN5.1 are contained in the subdirectory ./GEOScatchCNCLM51_GridComp.

Documentation for Catchment-CN5.1 was added here: [..]/GEOScatchCN_GridComp/GEOScatchCNCLM51_GridComp/README_CN51.md

Source code for Catchment-CN4.5 was removed throughout, except in the make_bcs package where the generation of model parameters for Catchment-CN4.5 is still needed for Catchment-CN5.1.

It is expected that the "remap_restarts" package is zero-diff for Catchment-CN4.0. There were, however, several bugs in CatchmentCNRst.F90 for Catchment-CN4.5 that were fixed for Catchment-CN5.1.

Related PRs (in the order in which they should be merged):

• GEOS_Util PR 149
• GEOSgcm_GridComp PR 1129 -- this PR
• GEOSgcm_App PR 751
• GEOSldas_GridComp PR 132
• GEOSldas PR 829

Outstanding issues to be resolved before merge:

• Add GEOSldas 0-diff test for CatchCN51.
• Standard 0-diff tests for GCM. (Completed by @biljanaorescanin on 7 Aug 2025.)
• Standard 0-diff tests for GEOSldas (Completed by @gmao-rreichle after commit 3438674.)
• 0-diff test for remapping Catchment-CN4.0 restart. (Completed by @biljanaorescanin on 7 Aug 2025.)
• Verify "remap_restarts" for Catchment-CN5.1. (Completed by @biljanaorescanin on 19 Aug 2025.)
• Add name of CTSM51 namelist file to GEOS_SurfaceGridComp.rc. (Verified by @biljanaorescanin on 18 Aug 2025.)
• Verify 0-diff vs. Jana/Eunjee global spin-up simulation. (Completed by @biljanaorescanin on 3 Sep 2025.)
• Make CatchCN51 work in the GCM. (Some of this is already in Add CatchmentCNCLM51 GEOSgcm_App#751.)
• Science verification of SIF output.
• Look into nitrogen imbalance. Warnings were turned off temporarily (67581a9; c742c1a). Could be caused by error in CNPFT restart variable indexing.
• Understand why nitrogen imbalance error statement is printed but run does not abort (see CatchCN CLM51 fix endrun exit issue #1152)
• Understand why T2MMIN5D values are seemingly wrong in restart file.
• Make sure calculations for T2MMIN5D are good. See inline comments Updates to add new Catchment-CN version: Catchment-CN5.1 #1129 (review).

To be addressed in a future PR:

• Improve efficiency:
  • Move computation of cumulative states and fluxes after call to catchcn(). See Updates to add new Catchment-CN version: Catchment-CN5.1 #1129 (comment).
  • Remove 90-minute sum ("m") InternalSpecs from catchcn restart file by (i) enforcing a simulation restart time and duration that is consistent with the 90-minute CN_Driver() time step and (ii) switching the order of the calls to catchment() and CN_Driver().

cc: @gmao-rreichle , @rdkoster , @weiyuan-jiang , @biljanaorescanin

[biljanaorescanin]

GCM 1 day AMIP, REPLAY and Inc Replay are zero diff.

[biljanaorescanin]

CatchCN CLM4.0 restart remapping compared from what I get using main GEOS_Util to what we get with GEOS-Util pr and this PR is zero diff for both m36 and c180 runs. Restart I used to remap from was one I created for Eunjee when she wanted to run v12 bcs.

Also, I forgot to say regression was also run as part of GCM testing on the exp PR and it was a PASS.

[biljanaorescanin]

Updates on my part:

1. Remap from CatchCN clm51 works.
2. I've played with the nml list, to rename it, change location and flag within the list and all works as intended.
   nml list echo is part of the log file so it is easy to double check options and confirm what is read.
3. I can't reproduce restart or output in general from my run compared to what Jana gave us to use RST files from.
4. I did prepare a "test" we can work with Matt to add to testing package.
5. I still see warnings for nitrogen imbalance in my ldas_err file.
   Lines like this: pe=00027 FAIL at line=00072 shr_abort_mod.F90 <ERROR in CNBalanceCheckMod.F90 at line 548>

[weiyuan-jiang]

I think the error is on this line

GEOSgcm_GridComp/GEOSagcm_GridComp/GEOSphysics_GridComp/GEOSsurface_GridComp/GEOSland_GridComp/Shared/lsm_routines.F90

Line 2728 in 6f95da3

integer, intent(out), optional :: rc ! needed for _ASSERT()

GNU fails to recognize that is a comment

[gmao-rreichle]

The treatment of T2MMIN5D doesn't seem right. See comments below

[gmao-rreichle]

The minimum air temperature used here is the "lowest-model-level" temperature, which is typically at about 40-60 m height, and not the screen-level (2m) temperature. There is a comment above (just before the if block) that suggests as much, but it's still misleading to label this variable T2m and claim as much in the LONG_NAME of the Internal_Spec definition.

[gmao-rreichle]

Since init_accum is hardwired to .false., "accper" here is undefined. This seems to be a bug.