Add MERRA2bias meteorological forcing reader and testcase

[Kristen-M-Whitney]

This pull request adds support for the MERRA2bias meteorological forcing
dataset within LIS.

The implementation builds upon the framework introduced in Issue #1760
(Add GEOS-IT bias meteorological forcing and extend MERRA-2 lapse rate
support), while adapting the reader for the MERRA2bias forcing dataset
used in the NLDAS-3 retrospective forcing workflow.

The MERRA2bias forcing dataset consists of native MERRA-2 near-surface
air temperature, near-surface specific humidity, and surface pressure,
together with bias-corrected downward longwave radiation generated
upstream of LIS.

The bias correction is applied using monthly MERRA-2 downward longwave
radiation climatologies, consistent with the approach used for the
CERES-based retrospective forcing record prior to 2026. Beginning in
2026, when CERES downward longwave radiation is no longer available,
the MERRA2bias dataset provides a continuous transition in the NLDAS-3
retrospective forcing workflow while preserving consistency with the
earlier CERES-based record.

Changes include:

• Added a new merra2bias meteorological forcing reader.
• Added support for MERRA2bias within the LIS metforcing plugin
  infrastructure.
• Added support for dynamic lapse-rate downscaling using MERRA-2-derived
  dynamic lapse-rate fields.
• Added documentation for the new forcing source.
• Added a functional testcase.

Functional testcase input files

Additional input files required to run the functional testcase are available at:

/discover/nobackup/projects/eis_freshwater/kwhitney/SHARE/merra2bias_testcase/

Resolves #1833

[emkemp]

Hi @Kristen-M-Whitney:

I'm attempting to run your merra2bias test case, but certain input data are missing and the location not reported in the README file. I'm specifically looking for:

./input/forcing_variables.txt
./input/MERRA2BIAS/
./input/dynlapserates_merra2bias/

Please update the README to provide information on where these files are.

[Kristen-M-Whitney]

@emkemp Thanks for catching that. Sorry, I forgot to note the location in the PR. I just edited the original PR to include the information. The additional input files are currently located at:

/discover/nobackup/projects/eis_freshwater/kwhitney/SHARE/merra2bias_testcase/

The files include those following the instructions: https://nasa-lis.github.io/LISF/howto_create_lis_testcases/howto_create_lis_testcases.html

I wasn't sure about adding this path to the README since it's an internal project directory and I wasn't sure whether everyone would have permission to access it. If you're unable to access the eis_freshwater project directory, let me know and I can relocate the files to my nobackup personal directory instead.

Would you prefer that I add the path to the README as well, or just keep it documented in the PR?

[emkemp]

@Kristen-M-Whitney @jvgeiger:

I took another crack at testing PR #1834, and I am having great difficulties in running LIS. I figured out I needed to change LD_LIBRARY_PATH and LIS_LIBESMF environment variables to use ESMF 8.6.1 installed by @jvgeiger. This fixes a problem with an undefined MPI_Communicator. But I'm now getting an MPI_Abort occurring as lisrun is called.

I'm assuming this is an EIS-Freshwater branch thing, which requires a special environment which I'm not very familiar with. Unless either of you can offer suggestions, I will punt this to @jvgeiger.

[dmocko]

Hi @emkemp

Yes, it may need the lisf_7.5_intel_2023.2.1_mapl module, which we use for eis-freshwater2 to
run the PFIO capability added there. Sorry for the trouble. Will give further details via e-mail.

[emkemp]

Update: @dmocko pointed me to a private module that includes MAPL. While the runtime behavior of LIS changes, it simply hangs now w/o opening a lislog file. I let a test run proceed for an hour in the debug QOS before it timed out, and the SLURM output file gave no useful information.

I'm punting.