PPE parameter work -Addresses issues 2830 and 2831 #2845

olyson · 2024-10-25T03:00:38Z

Description of changes

b4b moves of namelist variables to parameter file for PPE per issue #2830 and make some parameter variables pft-dimensioned per #2831

Note that baseflow_scalar and pot_hmn_ign_counts_alpha were not moved from namelist to parameter file due to reasons detailed in #2830

Specific notes

Contributors other than yourself, if any: @slevis-lmwg helped troubleshoot testing issues

CTSM Issues Fixed (include github issue #): #2830 , #2831

Are answers expected to change (and if so in what way)? No

Any User Interface Changes (namelist or namelist defaults changes)? Yes

Does this create a need to change or add documentation? Did you do so? No

Testing performed, if any:
Compare to ctsm5.3.009
Derecho aux_clm: PASS except for (not sure why the baseline files are missing, the last one is ok since it is a new test):

    FAIL ERP_D_P64x2_Ld3.f10_f10_mg37.I2000Clm50BgcCru.derecho_gnu.clm-default BASELINE ctsm5.3.009: ERROR BFAIL some baseline files were missing
    FAIL ERP_Ld5.f10_f10_mg37.I1850Clm50Bgc.derecho_gnu.clm-default BASELINE ctsm5.3.009: ERROR BFAIL some baseline files were missing
    FAIL ERS_D_Ld3.f10_f10_mg37.I1850Clm50BgcCrop.derecho_gnu.clm-default BASELINE ctsm5.3.009: ERROR BFAIL some baseline files were missing
    FAIL ERS_D_Ld6.f10_f10_mg37.I1850Clm45BgcCrop.derecho_gnu.clm-clm50CMIP6frc BASELINE ctsm5.3.009: ERROR BFAIL some baseline files were missing
    FAIL ERS_Ld3_D.f10_f10_mg37.I1850Clm50BgcCrop.derecho_gnu.clm-rad_hrly_light_res_half BASELINE ctsm5.3.009: ERROR BFAIL some baseline files were missing
    FAIL SMS_D.f10_f10_mg37.I2000Clm60Bgc.derecho_gnu BASELINE ctsm5.3.009: ERROR BFAIL some baseline files were missing
    FAIL SMS_Ln9.ne3pg3_ne3pg3_mg37.I2000Clm50Sp.derecho_gnu.clm-clm50cam6LndTuningMode--clm-nofireemis BASELINE ctsm5.3.009: ERROR BFAIL some baseline files were missing
    FAIL ERI_Ld9.f45_g37.I2000Clm50BgcCru.derecho_intel.clm-nofire BASELINE ctsm5.3.009: ERROR BFAIL some baseline files were missing
    FAIL ERS_D_Ld9.ne30pg3_t232.I1850Clm60BgcCropG.derecho_intel.clm-clm60cam7LndTuningMode BASELINE ctsm5.3.009: ERROR BFAIL some baseline files were missing
    FAIL SMS_Lm1.f10_f10_mg37.I1850Clm60Bgc.derecho_intel.clm-clm60_monthly_matrixcn_soilCN30 BASELINE ctsm5.3.009: ERROR BFAIL baseline directory '/glade/campaign/cgd/tss/ctsm_baselines/ctsm5.3.009/SMS_Lm1.f10_f10_mg37.I1850Clm60Bgc.derecho_intel.clm-clm60_monthly_matrixcn_soilCN30' does not exist

Izumi aux_clm: PASS except for (ok though):

PEND SMS_Ld10_D_Mmpi-serial.CLM_USRDAT.I1PtClm60Bgc.izumi_nag.clm-NEON-MOAB--clm-PRISM SHAREDLIB_BUILD (UNEXPECTED: expected FAIL)

nohup ./build-namelist_test.pl PASS
./run_ctsm_py_tests -u PASS (OK)
./run_ctsm_py_tests -s PASS (OK)

Jmaxb1 (now called jmaxb1 and pft-dependent) upplim_destruct_metamorph maximum_leaf_wetted_fraction interception_fraction

…monthly_matrixcn_soilCN30

slevis-lmwg

@olyson thank you for your work in this PR. My comments are mostly minor. The one about whether to read params in pftconMod or locally is a bit bigger, and I'm not certain whether it's relevant and worth considering.

slevis-lmwg · 2024-10-25T16:10:07Z

bld/namelist_files/namelist_defaults_ctsm.xml

+<paramfile phys="clm6_0" >/glade/campaign/cgd/tss/people/oleson/modify_param/ctsm60_params.c241017.nc</paramfile>
+<paramfile phys="clm5_1" >/glade/campaign/cgd/tss/people/oleson/modify_param/ctsm51_params.c241017.nc</paramfile>
+<paramfile phys="clm5_0" >/glade/campaign/cgd/tss/people/oleson/modify_param/clm50_params.c241017.nc</paramfile>
+<paramfile phys="clm4_5" >/glade/campaign/cgd/tss/people/oleson/modify_param/clm45_params.c241017.nc</paramfile>


@olyson at this point I will

copy the new files to .../lnd/clm2/paramdata and

./rimport them

...and you

can change the paths here.

slevis-lmwg · 2024-10-25T16:14:11Z

cime_config/testdefs/testlist_clm.xml

@@ -875,6 +875,15 @@
      <option name="comment"  >Run matrix solution fast spinup for restart beyond a year (to trigger next year logic) turning off extra options to make faster (so without crop or isotopes). This test would've caught a previous bug.</option>
    </options>
  </test>
+  <test name="SMS_Lm1" grid="f10_f10_mg37" compset="I1850Clm60Bgc" testmods="clm/clm60_monthly_matrixcn_soilCN30">


@olyson thank you for adding this back.

slevis-lmwg · 2024-10-25T16:15:16Z

cime_config/testdefs/testmods_dirs/clm/ciso_cwd_hr/user_nl_clm

@@ -1,2 +1,2 @@
-paramfile = '$DIN_LOC_ROOT/lnd/clm2/paramdata/ctsm51_ciso_cwd_hr_params.c240814.nc'
+paramfile = '/glade/campaign/cgd/tss/people/oleson/modify_param/ctsm51_ciso_cwd_hr_params.c241017.nc'


@slevis-lmwg will copy file and ./rimport it

@olyson will update the path

slevis-lmwg · 2024-10-25T16:16:02Z

cime_config/testdefs/testmods_dirs/clm/clm60_monthly_matrixcn_soilCN30/user_nl_clm

@@ -1,2 +1,2 @@
 use_soil_matrixcn = .true.
-paramfile = '$DIN_LOC_ROOT/lnd/clm2/paramdata/ctsm60_params_cn30.c240822.nc'
+paramfile = '/glade/campaign/cgd/tss/people/oleson/modify_param/ctsm60_params_cn30.c241017.nc'


@slevis-lmwg will copy file and ./rimport it

@olyson will update the path

slevis-lmwg · 2024-10-25T16:22:32Z

src/biogeochem/CNGapMortalityMod.F90

@@ -216,12 +209,12 @@ subroutine CNGapMortality (bounds, num_soilp, filter_soilp, &
               am = min(1._r8, am + heatstress(p))
            else ! lpj didn't set this for grasses; cn does
               ! set the mortality rate based on annual rate
-               am = params_inst%am
+               am = r_mort(ivt(p))


I saw that you removed "am" from line 37 above. That makes me wonder whether line 212 -- and the "am = min(..." three lines above -- are still used somewhere?

That is a good point. The variable am is within a use_cndv if block which is now deprecated. I don't think we have any CNDV tests anymore, but if we did I think we wouldn't get the same answers as before with my changes here because "m" would not longer be multiplied by "am". In theory we could remove that whole section of code. But if we don't want to remove it for some reason. I guess I could change the code to this, so that if use_cndv is invoked, then am will be used, as it was before. But we don't have any CNDV testing to check this.

do fp = 1,num_soilp p = filter_soilp(fp) if (use_cndv) then ! Stress mortality from lpj's subr Mortality. if (woody(ivt(p)) == 1._r8) then if (ivt(p) == 8) then mort_max = 0.03_r8 ! BDT boreal else mort_max = 0.01_r8 ! original value for all patches end if ! heatstress and greffic calculated in Establishment once/yr ! Mortality rate inversely related to growth efficiency ! (Prentice et al 1993) am = mort_max / (1._r8 + k_mort * greffic(p)) ! Mortality rate inversely related to growth efficiency ! (Prentice et al 1993) am = mort_max / (1._r8 + k_mort * greffic(p)) am = min(1._r8, am + heatstress(p)) else ! lpj didn't set this for grasses; cn does ! set the mortality rate based on annual rate am = r_mort(ivt(p)) end if else am = r_mort(ivt(p)) end if m = am/(get_average_days_per_year() * secspday)

As far as I know, cndv has not worked (or not worked correctly, at least) for quite some time. So I would be fine with removing cndv code. I don't know if @ekluzek would agree.

Thanks @slevis-lmwg . I'm thinking now that maybe it wouldn't be a good idea to remove the cndv code piecemeal. I think for now I'll make the code mod I suggested above.

slevis-lmwg · 2024-10-25T16:52:09Z

src/main/pftconMod.F90

+     real(r8), allocatable :: jmaxb0(:)           ! Baseline proportion of nitrogen allocated for electron transport (J)
+     real(r8), allocatable :: jmaxb1(:)           ! Coefficient determining the response of electron transport rate to light availability (-)
+     real(r8), allocatable :: wc2wjb0(:)          ! The baseline ratio of rubisco limited rate vs light limited photosynthetic rate (Wc:Wj) (-)
+     real(r8), allocatable :: r_mort(:)           ! Mortality rate (1/year)


@olyson I understand that the code works as written. I will ask this anyway in case it's relevant and worth considering:

As a general rule we try to place here (pftconMod) the params that appear in multiple places in the code. On the other hand, we try to place params that appear in a single module in the module where they appear.

Right, I started down the path of placing the parameters within LunaMod, but realized we don't have the infrastructure in there to do that (e.g., as is done in PhotosynthesisMod.F90 with the photo_params_type and subroutines allocParams, cleanParams, etc.). Much easier to put them in pftconMod.F90 instead of putting all of that infrastructure in, and I wasn't sure what the real value of all that was anyway. But if there is consensus to do it that way, I can do it.

I am also often motivated by efficiency in code development, so I will let @ekluzek have final say about this.

Thinking about it some more, now that my head is in parameter space, I think I'll just put the work in now to have the parameters read in in LunaMod.F90.

And CNGapMortalityMod.F90 (for r_mort).

…to CNGapMortality

olyson · 2024-10-26T18:29:45Z

I think this is ready for re-review. Thanks for the input @slevis-lmwg .
I've moved the read of jmaxb0, jmaxb1, wc2wjb0 from pftconMod to LunaMod and r_mort to CNGapMortality.
Also, made adjustment to r_mort for use_cndv.
Testing results are exactly the same as reported previously.
I have not changed the paths to the parameter files yet in case there are more changes and testing.

olyson · 2024-10-27T20:19:31Z

I see now that @slevis-lmwg has rimported the parameter files so I have changed the paths accordingly.
A question I have is that I see this when I ran build-namelist_text.pl:

Looks like you planned 3997 tests but ran 3996.

Does this have anything to do with the jmax test I removed? Is there something else I should remove?

olyson · 2024-10-27T20:28:12Z

@slevis-lmwg , it looks like permissions need to be changed on those parameter files, thanks.

olyson added 8 commits October 20, 2024 10:09

Transfer the following parameters from namelist to parameter file:

694333f

Jmaxb1 (now called jmaxb1 and pft-dependent) upplim_destruct_metamorph maximum_leaf_wetted_fraction interception_fraction

Make jmaxb0, wc2wjb0, r_mort pft-dimensioned

77b8e80

ncd_io not needed after all

b7121a6

Other variables not needed after all

3f4936c

Consistent naming of units

bc3d518

New versions of specialized parameter files

e5100a4

Reinstate SMS_Lm1.f10_f10_mg37.I1850Clm60Bgc.derecho_intel.clm-clm60_…

28d552e

…monthly_matrixcn_soilCN30

Merge branch 'b4b-dev' into b4b-dev_PPE2830-2831

f91cd8a

olyson added the bfb bit-for-bit label Oct 25, 2024

olyson requested a review from slevis-lmwg October 25, 2024 03:00

slevis-lmwg assigned olyson Oct 25, 2024

slevis-lmwg requested changes Oct 25, 2024

View reviewed changes

Move read of jmaxb0,jmaxb1,wc2wjb0 from pftcon to LunaMod and r_mort …

1571076

…to CNGapMortality

Final paths for parameter files

0fd1a04

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

PPE parameter work -Addresses issues 2830 and 2831 #2845

PPE parameter work -Addresses issues 2830 and 2831 #2845

olyson commented Oct 25, 2024

slevis-lmwg left a comment

slevis-lmwg Oct 25, 2024 •

edited

Loading

slevis-lmwg Oct 25, 2024

slevis-lmwg Oct 25, 2024 •

edited

Loading

slevis-lmwg Oct 25, 2024 •

edited

Loading

slevis-lmwg Oct 25, 2024

olyson Oct 25, 2024

slevis-lmwg Oct 25, 2024

olyson Oct 25, 2024

slevis-lmwg Oct 25, 2024

olyson Oct 25, 2024

slevis-lmwg Oct 25, 2024

olyson Oct 25, 2024

olyson Oct 25, 2024

olyson commented Oct 26, 2024

olyson commented Oct 27, 2024

olyson commented Oct 27, 2024

		@@ -1,2 +1,2 @@
		paramfile = '$DIN_LOC_ROOT/lnd/clm2/paramdata/ctsm51_ciso_cwd_hr_params.c240814.nc'
		paramfile = '/glade/campaign/cgd/tss/people/oleson/modify_param/ctsm51_ciso_cwd_hr_params.c241017.nc'

PPE parameter work -Addresses issues 2830 and 2831 #2845

Are you sure you want to change the base?

PPE parameter work -Addresses issues 2830 and 2831 #2845

Conversation

olyson commented Oct 25, 2024

Description of changes

Specific notes

slevis-lmwg left a comment

Choose a reason for hiding this comment

slevis-lmwg Oct 25, 2024 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

slevis-lmwg Oct 25, 2024 • edited Loading

Choose a reason for hiding this comment

slevis-lmwg Oct 25, 2024 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

olyson commented Oct 26, 2024

olyson commented Oct 27, 2024

Looks like you planned 3997 tests but ran 3996.

olyson commented Oct 27, 2024

slevis-lmwg Oct 25, 2024 •

edited

Loading

slevis-lmwg Oct 25, 2024 •

edited

Loading

slevis-lmwg Oct 25, 2024 •

edited

Loading