H20 fix for historical.

[mvdebolskiy]

@JessicaNeedham can you describe how you have tested this? And some reasoning.

[JessicaNeedham]

No regression testing.

The original historic simulation crashed in year 2001. I branched off that run from a restart in 1994 (new case using finidat) with this change included and it ran to 2016 with no crashes. Case on betzy here: /cluster/work/users/jessica/ncsrevise_runs/noresm-fates-dumpsite_terminate-cohorts_history.2026-05-27

I ran the diagnostics package for the overlapping years (1994-2000) and results are here: https://ns9560k.web.sigma2.no/datalake/diagnostics/noresm/jessica/noresm-fates-dumpsite_terminate-cohorts_history.2026-05-27/compare/noresm-fates-ne16-LU-PPE-1901-2024.2026-05-03/ANN/

These results look reasonable - very small changes in FATES_AREA_PLANTS and FATES_GRAZING - consistent with the termination of these small grassy cohorts.

Reasoning
As @rosiealice points out in #218 - it looks like what is happening is that grazing is removing the leaves of grasses and the fine roots are decaying due to lack of NPP. Since we have no sapwood respiration, there are no respiring tissues and so these cohorts don't decrease stored carbon and don't die of carbon starvation mortality. They remain as immortal stalks. These changes add a new termination mortality term that removes evergreen, grassy cohorts that have no leaf or fineroot biomass.

A long term solution would be to turn on sapwood respiration or to allow grazing to reduce storage carbon. I tried adding grazing of storage but restarting the run with that in 1850 resulted in a mass balance error.

[JessicaNeedham]

Note - the run that crashed and the branch off it with the fix are both prior to Beta16. I have not tested with Beta16 code.

[mvdebolskiy]

a mass balance error
That is most likely because of the timing or you have not put the update to the correct routine. They work differently on the restart. Can you make a PR with that change, I might have a quick fix.

[JessicaNeedham]

a mass balance error
That is most likely because of the timing or you have not put the update to the correct routine. They work differently on the restart. Can you make a PR with that change, I might have a quick fix.

PR here: #62