[billsacks]

Description of changes

Add an is_beg_curr_year conditional around file-based land cover change.

In most situations, this extra check of is_beg_curr_year should be
redundant with the way this land cover change was implemented. However,
it serves two purposes:

(1) Avoiding unnecessary work

(2) For a startup / hybrid run that starts mid-year, in some situations
we can get area changes in the first time step of the run without
this is_beg_curr_year check. This causes an error in annual flux
dribblers that are told to only expect fluxes at the year boundary.
This is_beg_curr_year suppresses any area changes until the first
year boundary of the run.

The main motivation for this change was to resolve (2). This resolves #4165.

I have also removed comments describing how to change the area change
behavior to be time-interpolated rather than uninterpolated. I don't
think this has actually been easy to change for a long time due to the
use of annual flux dribblers. And the introduction of is_beg_curr_year
even further bakes in the current behavior. However, I now discuss this in a separate design doc, also part of this PR.

Specific notes

CTSM issues resolved or otherwise addressed, if any:

• Resolves Mid-year startup with DGLC or CISM can lead to unexpected non-zero delta #4165

If answers are expected to change, describe (delete this line otherwise):

• Changes answers for some tests that start mid-year. However, answer changes only appear in two situations: (1) cases that would have failed before this change, or (2) cases with the for_testing_allow_non_annual_changes namelist variable set to true. In the aux_clm test suite, this only shows up in two tests: ERI_D_Ld20.f10_f10_mg37.IHistClm60Bgc.derecho_intel.clm-default--clm-midDecStart--clm-allowTransientYearJump and ERI_D_Ld20.f10_f10_mg37.IHistClm60Bgc.izumi_nag.clm-default--clm-midDecStart--clm-allowTransientYearJump.

Any user interface changes (namelist or namelist defaults changes)? No (but the allowTransientYearJump testmod no longer sets for_testing_allow_non_annual_changes).

Testing planned or performed, if any:

• aux_clm: tests pass and are bit-for-bit except as noted above (a few tests were still pending when I opened this PR)

Requirements before merge:

• I have followed the CTSM contribution guidelines.

• The code in this PR branch builds with no errors.
• The code in this PR branch runs with no errors. Briefly describe tested configuration(s): full aux_clm
• This either (a) does not change answers, (b) it only changes answers at roundoff level, or (c) I have performed a scientific evaluation of the answer changes. Which?: Mostly (a), with a minor exception as noted above

• I have reviewed relevant parts of the CLM documentation Tech Note or User's Guide to determine if anything needs to be changed or added. If it does, describe:
• This PR either (a) does not create a need to update the documentation or (b) includes required documentation updates (see guidelines for contributing documentation). Which?: (a) does not need an update to the documentation

[billsacks]

I struggled for a while with how to fix #4165 without causing at least as many issues as I was resolving. This is the best fix I could come up with. As noted in the accompanying design doc, it further bakes in some current behavior, which I don't love, but I think this is acceptable for now. If we ever want to loosen these restrictions, we could remove the check for non-annual changes from the dribblers, which would be another way to resolve #4165 so would allow removing this added is_beg_curr_year check.

Note that the change here changes behavior for a run starting mid-year... but in the situations where it changes behavior, previously you would have needed for_testing_allow_non_annual_changes to avoid an abort, so this change seems like it should be fine / good.

It's probably best if this PR comes in for CESM3 because without it users could run into errors when trying to start a run mid-year.