Disable hydraulic failure mortality for leafless plants

mpaiao commented 11 months ago

Description:

These changes are mostly to be on the safe side. Hydraulic failure mortality is triggered when btran is too low (non-hydro mode) or when hydraulic conductivity is diminished (hydro mode). This pull request adds a check to see if the PFT is deciduous and is completely leafless. In these cases, hydraulic failure is forced to zero.

This change impacts both drought-deciduous and cold-deciduous PFTs.
In the future, we may want to coordinate hydraulic failure with phenology, to ensure that deciduous plants abscise leaves instead of dying when under high risk of hydraulic failure.

Collaborators:

Expectation of Answer Changes:

This may have minimal effect under severely dry conditions.

Checklist

If this is your first time contributing, please read the CONTRIBUTING document.

All checklist items must be checked to enable merging this pull request:

Contributor

[x] The in-code documentation has been updated with descriptive comments
[ ] The documentation has been assessed to determine if updates are necessary

Integrator

[ ] FATES PASS/FAIL regression tests were run
[ ] Evaluation of test results for answer changes was performed and results provided

Documentation

Technical Note update:
User's Guide update:

Test Results:

CTSM (or) E3SM (specify which) test hash-tag:

CTSM (or) E3SM (specify which) baseline hash-tag:

FATES baseline hash-tag:

Test Output:

mpaiao commented 10 months ago

Thanks for the feedback @ZacharyRobbins! I'm not very familiar with FATES-Hydro, but if the current FATES-Hydro implementation allows for plants to continue losing water when they are leafless, then we definitely don't want this change, at least when FATES-Hydro is on. Let me know and I can update/close the PR.

rosiealice commented 10 months ago

So the loss of water from leafless plants is a big issue for frozen soils, which all else being equal, dehydrate the plants via the roots, leading to massive hydraulic failure.

To prevent this, in Marius Lambert's work, the (minimum) stomatal conductance and the root conductivity needed to be dialled down (a lot) as a function of hardening (from his earlier paper).

Unfortunately this didn't make it to the trunk (Marius started working on river DOC immediately after finishing his thesis), and we are writing a proposal to follow it up.

But the lessons learned remain valid, and I think maybe transferable to the super dry soil conditions also where we know that Piñon pines become hydraulically isolated when swp <-3Mpa, from all of the work in NM etc.

If we turn this option on in fates hydro it might help stop things dying in the off season (if that is desired) but there would be a potentially big pulse of mortality when the leaves came back on... we could prevent that by stopping the leaves coming in if the plc was high, but it would probably be better to stop the (bulk of the) dehydration process in the first place?

mpaiao commented 10 months ago

Thanks for the feedback @rosiealice, and good to know that there is some more mechanistic approach to prevent the mortality. So maybe for now the best option is to implement my original suggestion only when FATES-Hydro is turned off, and note that in the future we will incorporate Marius' work?

ZacharyRobbins commented 10 months ago

I think that makes sense @mpaiao .

glemieux commented 10 months ago

@JessicaNeedham and @mpaiao, in deconflicting main against this PR branch, I was unsure how to reconcile the update from https://github.com/NGEET/fates/pull/1115. The current code on main has the following: https://github.com/NGEET/fates/blob/a9bbfd820bb613afd0c33fe75404c7848b435260/biogeochem/EDMortalityFunctionsMod.F90#L159-L162

If we have time, we can discuss together at the fates software meeting today.

mpaiao commented 10 months ago

@glemieux Thanks for pointing out the conflict. Perhaps the simplest way would be to rewrite the if block as the following:

          ! When FATES-Hydro is off, hydraulic failure mortality occurs only when btran
          ! falls below a threshold, soils are not frozen and plants have leaves.
          if ( ( .not. is_decid_dormant ) .and. & 
               ( btran_ft(cohort_in%pft) <= hf_sm_threshold ) .and. &
               ( ( minval(bc_in%t_soisno_sl) - tfrz ) > soil_tfrz_thresh ) ) then
             hmort = EDPftvarcon_inst%mort_scalar_hydrfailure(cohort_in%pft)
          else
             hmort = 0.0_r8
          end if

glemieux commented 10 months ago

Regression testing on derecho is complete. All expected tests are B4B against the latest tag baseline with the exception of one long term test. DIFFs are not unexpected, but I'm going to look into these briefly before next steps.

location: /glade/u/home/glemieux/scratch/ctsm-tests/tests_pr1130-fates

glemieux commented 10 months ago

This DIFF starts up 2.5 months into the 6 month tests and is occurring only on one gridcell on the coast of Antarctica for that whole time. As such, I think this is good to integrate.

NGEET / fates