Add some biophysical variables by land-use type #1407
Conversation
github-project-automation
bot
moved this to Finding Reviewers
in FATES Pull Request Planning and Status
ckoven
added
type: enhancement
software: history output
|
|
||
| ! for the rest of these, first weight by the vegetated area of each patch over the total patch area for each land use type | ||
| hio_tsa_si_landuse(io_si,cpatch%land_use_label) = hio_tsa_si_landuse(io_si,cpatch%land_use_label) + & | ||
| bc_in(s)%t2m_pa(cpatch%patchno) * cpatch%total_canopy_area/landuse_statevector(cpatch%land_use_label) |
There was a problem hiding this comment.
For normalization factors like cpatch%total_canopy_area/landuse_statevector(cpatch%land_use_label) that are repeated a bunch, it'd be good to create a variable for them with a useful name. E.g., what I did in update_history_dyn_subsite_ageclass() with patch_area_div_site_area.
| cpatch => cpatch%younger | ||
| end do | ||
|
|
||
| ! for all the land-use indexed variables, excpet for TVEG, also add in the component for the unvegetated area of each land use |
| real(r8) :: landuse_statevector(n_landuse_cats) | ||
| real(r8) :: canopy_area_bylanduse(n_landuse_cats) | ||
| integer :: i_lu | ||
| logical :: foundbaregroundpatch |
There was a problem hiding this comment.
foundbaregroundpatch appears to be set but unused.
| if (cpatch%land_use_label .eq. nocomp_bareground_land .and. .not. foundbaregroundpatch) then | ||
| foundbaregroundpatch = .true. | ||
| do i_lu = 1, n_landuse_cats | ||
| if ( landuse_statevector(i_lu) .gt. rsnbl_math_prec ) then |
There was a problem hiding this comment.
To reduce indentation and thus readability, this if ( landuse_statevector(i_lu) .gt. rsnbl_math_prec ) then ... could be replaced by cycling on its inverse—if ( landuse_statevector(i_lu) .le. rsnbl_math_prec ) then cycle.
|
|
||
|
|
||
| cpatch => sites(s)%oldest_patch | ||
| do while(associated(cpatch)) |
There was a problem hiding this comment.
Why is this do loop separate from the others? Is it because GPP is the only one with the ccohort%n * dt_tstep_inv normalization factor? A comment would be useful to explain this.
| ccohort => cpatch%shortest | ||
| do while(associated(ccohort)) | ||
| if ( .not. ccohort%isnew ) then | ||
| if (cpatch%land_use_label .gt. nocomp_bareground_land) then |
There was a problem hiding this comment.
This if statement could be combined with the one above to reduce indentation.
| ivar=ivar, initialize=initialize_variables, index = ih_lhflux_si_landuse ) | ||
|
|
||
| call this%set_history_var(vname='FATES_GPP_LU', units='kg m-2 s-1', & | ||
| long='gross primary productivity by age bin in kg carbon per m2 per second', & |
There was a problem hiding this comment.
"age bin" should be "land use type".
github-project-automation
bot
moved this from Finding Reviewers
to Under Review
in FATES Pull Request Planning and Status
|
@ckoven where is this on the priority queue for the CLM6 code freeze? |
glemieux
added
the
software: API
In order to look at biophysical effects of land use, it's useful to have a small number of relevant variables indexed by land use type. This PR allows that, so that you can look at: vegetation temperature, surface air temperature, and a basic set of energy fluxes (SW, LW, SH, LH) separately for each land use type. I.e., they are output on the
fates_levlandusehistory dimension for each gridcell.This requires HLM-side logic. I have a commit on E3SM that handles this at ckoven/E3SM@5fd137b (PR forthcoming)
Currently these are all output when the hi frequency history flag is set to 1; in principle this should actually be when the flag is set to 2, so that should probably change before bringing this PR in. (for the purpose of the experiments I am currently running, I want these and only these subgridscale high-frequency variables though.)
Also, the logic that averages in bare-ground-patch properties will only work for nocomp configurations. Further work is needed to be correct for full-FATES configurations as well, but I got lost in how to index a given site's bare-ground patch when nocomp isn't on, so I need to come back to that.
Also, the patch-weighting logic for these variables is slightly different than for other patch-level variables (e.g., the `fates_levage' -indexed ones). This is for two reasons: (1) we need to average in some of the bare ground patch properties to weight things correctly, because the FATES and HLM patch areas mean different things, and (2) the land use areas are less generally dynamic than the age binned logic, and so there is less chance of weirdness when the averaging denominator changes abruptly. So the upshot is that less postprocessing is needed for these (i.e they are simpler to work with), but also that they might not sum perfectly to the total site values because of changing denominators over time.
Description:
fixes #1404.
Collaborators:
discussed with @glemieux, @rgknox, @aswann
Expectation of Answer Changes:
This only adds new diagnostic variables, so should be bit for bit otherwise.
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
Integrator
If satellite phenology regressions are not b4b, please hold merge and notify the FATES development team.
Documentation
Test Results:
CTSM (or) E3SM (specify which) test hash-tag:
CTSM (or) E3SM (specify which) baseline hash-tag:
FATES baseline hash-tag:
Test Output: