Handmerge feature/pcolarco/#1-gocart2g-updates into develop #68

mathomp4 · 2021-08-04T16:16:34Z

This is a PR to merge in feature/pcolarco/#1-gocart2g-updates into develop. This was done as a hand-merge to be careful about it (rather than depending on GitHub).

Note, this possibly could affect the UFS work so I'm pinging @rmontuoro as this is what we want all future development based on. I think I kept things working so that your UFS code still works.

As far as I know how to test GOCART2G this is zero diff to feature/pcolarco/#1-gocart2g-updates. I did a c24 run with OPS GOCART2G in 2015 and...it was zero-diff?

For @pcolarco's benefit, develop "won" in the merge only in a few places as seen here:

https://github.com/GEOS-ESM/GOCART/compare/feature/pcolarco/%231-gocart2g-updates..feature/mathomp4/merge-develop-pcolarco-%231-gocart2g-updates

One main change that was on develop is that the scheme in dust is now specified by a string and not an integer. We can always undo this as needed, but it allowed for an easier merge. So:

emission_scheme: 1 # Ginoux

becomes:

emission_scheme: ginoux

The other main change is that in the Dust State Specs we went from:

 DU_SAND    | 1      | xy   | N   |      | sand_fraction
 DU_SILT    | 1      | xy   | N   |      | silt_fraction
 DU_CLAY    | 1      | xy   | N   |      | clay_fraction

to:

 DU_SAND    | 1      | xy   | N   |      | volume_fraction_of_sand_in_soil
 DU_SILT    | 1      | xy   | N   |      | volume_fraction_of_silt_in_soil
 DU_CLAY    | 1      | xy   | N   |      | volume_fraction_of_clay_in_soil

I think this is because @rmontuoro and UFS need a CF compliant standard name in the LONG NAME and not our...not so compilant names.

…from NIthermo.

removed diagnostic print statements

…d is in Chem_AeroGeneric

include BR

mathomp4 · 2021-08-04T16:42:57Z

@rmontuoro It would be good if you can test this branch with UFS and make sure nothing got broken on your end.

rmontuoro

These change will break UFS-Aerosols and require further code development on the NOAA side.

rmontuoro · 2021-08-04T17:02:37Z

ESMF/GOCART2G_GridComp/DU2G_GridComp/DU2G_StateSpecs.rc

- DU_CLAY    | 1      | xy   | N   | scheme == 'fengsha' | volume_fraction_of_clay_in_soil
- DU_SAND    | 1      | xy   | N   | scheme == 'fengsha' | volume_fraction_of_sand_in_soil
- DU_SILT    | 1      | xy   | N   | scheme == 'fengsha' | volume_fraction_of_silt_in_soil
+ DU_SRC     | 1      | xy   | N   |      | erod - dust emissions


These DU_* fields should be associated with a conditional statement (scheme == name) according to early discussions with @tclune.
Is there a particular reason why the COND column should be empty in this update?

@rmontuoro Well how do you do it if there are two schemes that need it? Both k14 and ginoux use du_src? I removed it because of that.

As for the others, DU_Z0, DU_GVF are k14 only, so those could be scheme == 'k14' protected.

But DU_SAND, DU_SILT, and DU_CLAY look to be used by both k14 and fengsha.

@mathomp4 - thanks for the explanation. This is a question for @tclune. I believe something like:

scheme == 'ginoux' .or. scheme == 'k14'

could work.

Okay, it does look like scheme == 'ginoux' .OR. scheme == 'k14' could work here. So maybe we can protect the IMPORT pointers that are affecting you a la:

DU_SRC | 1 | xy | N | scheme == 'ginoux' .OR. scheme == 'k14' | erod - dust emissions DU_Z0 | 1 | xy | N | scheme == 'k14' | aerodynamic_surface_roughness_for_aeolian_processes DU_GVF | 1 | xy | N | scheme == 'k14' | GVF DU_SAND | 1 | xy | N | scheme == 'fengsha' .OR. scheme == 'k14' | volume_fraction_of_sand_in_soil DU_SILT | 1 | xy | N | scheme == 'fengsha' .OR. scheme == 'k14' | volume_fraction_of_silt_in_soil DU_CLAY | 1 | xy | N | scheme == 'fengsha' .OR. scheme == 'k14' | volume_fraction_of_clay_in_soil DU_RDRAG | m-1 | xy | N | scheme == 'fengsha' | drag_partition DU_SSM | 1 | xy | N | scheme == 'fengsha' | sediment_supply_map DU_UTHRES | m s-1 | xy | N | scheme == 'fengsha' | surface_dry_threshold_velocity FRSNOW | 1 | xy | N | scheme == 'fengsha' | surface_snow_area_fraction SLC | 1 | xy | N | scheme == 'fengsha' | liquid_water_content_of_soil_layer DU_TEXTURE | 1 | xy | N | scheme == 'k14' | soil_texture DU_VEG | 1 | xy | N | scheme == 'k14' | vegetation_type ... U10N | m s-1 | xy | N | scheme == 'k14' | equivalent_neutral_10-meter_eastward_wind V10N | m s-1 | xy | N | scheme == 'k14' | equivalent_neutral_10-meter_northward_wind ... WCSF | m3 m-3 | xy | N | scheme == 'k14' | water_surface_layer

Though TROPP is harder as it is scheme agnostic.

That's doable:

category: IMPORT #---------------------------------------------------------------------------------------- # VARIABLE | DIMENSIONS | Additional Metadata #---------------------------------------------------------------------------------------- NAME | UNITS | DIMS | VLOC | COND | LONG NAME #---------------------------------------------------------------------------------------- DU_SRC | 1 | xy | N | scheme == 'ginoux' .OR. scheme == 'k14' | erod - dust emissions DU_Z0 | 1 | xy | N | scheme == 'k14' | aerodynamic_surface_roughness_for_aeolian_processes DU_GVF | 1 | xy | N | scheme == 'k14' | GVF DU_SAND | 1 | xy | N | scheme == 'fengsha' .OR. scheme == 'k14' | volume_fraction_of_sand_in_soil DU_SILT | 1 | xy | N | scheme == 'fengsha' .OR. scheme == 'k14' | volume_fraction_of_silt_in_soil ...

@rmontuoro I can supply you with this file if you like so you can test with it before I push to this branch?

@mathomp4 - sure.

@rmontuoro here you go. I had to add ".txt" at the end so GitHub would upload it, so you'll have to un-do that.

DU2G_StateSpecs.rc.txt

Thanks @mathomp4. I've built and run this update within the UFS as a preliminary test after some integration work, and suggested a few changes in my latest code review.

I would also like to add the changes in my PR #65 as it looks like they are still relevant.

rmontuoro · 2021-08-04T17:19:07Z

ESMF/GOCART2G_GridComp/DU2G_GridComp/DU2G_StateSpecs.rc

 WET1       | 1      | xy   | N   |      | surface_soil_wetness
 LWI        | 1      | xy   | N   |      | land-ocean-ice_mask
- FRLAKE     | 1      | xy   | N   |      | fraction_of_lake
+ TROPP      | Pa     | xy   | N   |      | tropopause_pressure_based_on_blended_estimate


These new imports are concerning. They will break the UFS model and require further development work that cannot occur at this time, since we are just starting our evaluation runs.

@pcolarco - could you please clarify how satisfy/compute the following imports?

TROPP

U10N/V10N

WCSF

Of these, U10N, V10N, and WCSF seem to be K14 related.

TROPP though, seems to be GEOS endemic and it seems to come from tropovars in https://github.com/GEOS-ESM/GMAO_Shared/blob/main/GEOS_Shared/tropovars.F90. I think TropP3 is eventually TROPP_BLENDED

@mathomp4 - Thank you, this is helpful. We could implement tropovars in UFS-Aerosols to compute TROPP inline, but it requires potential vorticity as an input and this field is not available at this time.

Hmm. Well, you are beyond my ken. Maybe @amdasilva or someone with science knowledge could help come up with a tropopause pressure that could be used as an approximate proxy? (Also, I have no idea if that is actually where TROPP_BLENDED comes from, but I think it might from my search.)

If I understand it correctly, a "blended" tropopause pressure is determined "blending" the tropopause k-levels estimated by looking at temperature (ktherm) and potential vorticity (kept) profiles:

kblend = max( ktherm,kepv )

We could use other methods to compute the pressure at the tropopause that do not require the potential vorticity (see, for example, Reichler et al., 2003). However, I don't know how they would compare with the blended method and this seems to introduce an uncomfortably high degree of uncertainty at this stage.

Definitely needs a comment from @amdasilva

@rmontuoro There is another possibility: maybe you don't care? If I look at Process_Library/GOCART2G_Process.F90, tropp is only used to calculate stexttau and stscatau, e.g.,:

GOCART/Process_Library/GOCART2G_Process.F90

Lines 3585 to 3591 in c66c311

if( present(stexttau) .and. associated(stexttau) ) then

if (ple(i,j,k) .le. tropp(i,j)) then

stexttau(i,j,w) = stexttau(i,j,w) + tau

elseif(ple(i,j,k) .gt. tropp(i,j) .and. ple(i,j,k-1) .lt. tropp(i,j)) then

stexttau(i,j,w) = stexttau(i,j,w) + log(tropp(i,j)/ple(i,j,k-1))/log(ple(i,j,k)/ple(i,j,k-1))*tau

endif

endif

So, if UFS doesn't need to know about "stratospheric ext. AOT at 550 nm" or "stratospheric sct. AOT at 550 nm" maybe it's a throwaway? Just fill it with something that won't cause a "divide-by-zero" and move along?

@mathomp4 - good catch! It is hard to make a decision on this right now since we are still discussing how aerosol-radiation feedback will be introduced in the UFS using GOCART. We might need to provide stexttau to the radiation scheme running in the atmospheric model. Also, I would like to preserve the ability to compute these diagnostics in UFS-Aerosols. Nevertheless, your suggestion is good and could be a starting point.
This also means that the accuracy of TROPP may not play such a big role here.

One other option would be to add a procedure computing the atmospheric pressure at the tropopause to the process library. This would help achieve consistency across different applications.

tclune · 2021-08-05T12:07:14Z

ESMF/GOCART2G_GridComp/CA2G_GridComp/CCMI_REF-C2/CA2G_instance_CA.br.rc

@@ -26,7 +21,7 @@ hydrophobic_fraction: 0.5
 fscav: 0.0  0.4

 # particle radius
-particle_radius_microns: 0.0212  0.0212
+particle_radius_microns: 0.35 0.35


Some of the shifts in parameter values are disconcerting. This one in particular which is > 10x.

tclune

Well, its different. Hmm.

mathomp4 · 2021-08-05T15:37:41Z

Okay. I pushed up the latest Dust State Specs with the scheme logic as it seems to be zero-diff for ginoux and k14. I also added back a print I mistakenly removed along with an assert in SetServices (to make sure you supply a supported scheme), and put in the three fengsha settings that were in develop.

Of course, fengsha fails. I tried it with the ginoux settings and you of course get:

        EXTDATA: INFO: ExtData could not satisfy import DU_RDRAG
        EXTDATA: INFO: ExtData could not satisfy import DU_SSM
        EXTDATA: INFO: ExtData could not satisfy import DU_UTHRES
        EXTDATA: INFO: ExtData could not satisfy import FRSNOW
        EXTDATA: INFO: ExtData could not satisfy import SLC
pe=00080 FAIL at line=00851    MAPL_ExtDataGridCompMod.F90              <Found   5 unfullfilled imports in extdat>

because we don't have all the necessary imports in GEOS yet. (And I need to fix that error message in MAPL.)

rmontuoro

Additional changes are required to make this update work with the UFS.

rmontuoro · 2021-08-06T00:51:30Z

ESMF/GOCART2G_GridComp/DU2G_GridComp/DU2G_GridCompMod.F90

@@ -150,8 +147,20 @@ subroutine SetServices (GC, RC)
    call ESMF_ConfigGetAttribute (cfg, self%Ch_DU_res,  label='Ch_DU:', __RC__)
    call ESMF_ConfigGetAttribute (cfg, self%rlow,       label='radius_lower:', __RC__)
    call ESMF_ConfigGetAttribute (cfg, self%rup,        label='radius_upper:', __RC__)
+    call ESMF_ConfigGetAttribute (cfg, self%clayFlag,   label='clayFlag:', __RC__)


The statements on line 150-153 need to be guarded as they are only required for the k14 dust scheme.
I suggest consolidating the _ASSERT on line 159 in a select case block. See patch below:

--- a/ESMF/GOCART2G_GridComp/DU2G_GridComp/DU2G_GridCompMod.F90 +++ b/ESMF/GOCART2G_GridComp/DU2G_GridComp/DU2G_GridCompMod.F90 @@ -147,10 +147,6 @@ contains call ESMF_ConfigGetAttribute (cfg, self%Ch_DU_res, label='Ch_DU:', __RC__) call ESMF_ConfigGetAttribute (cfg, self%rlow, label='radius_lower:', __RC__) call ESMF_ConfigGetAttribute (cfg, self%rup, label='radius_upper:', __RC__) - call ESMF_ConfigGetAttribute (cfg, self%clayFlag, label='clayFlag:', __RC__) - call ESMF_ConfigGetAttribute (cfg, self%f_swc, label='soil_moisture_factor:', __RC__) - call ESMF_ConfigGetAttribute (cfg, self%f_scl, label='soil_clay_factor:', __RC__) - call ESMF_ConfigGetAttribute (cfg, self%uts_gamma, label='uts_gamma:', __RC__) call ESMF_ConfigGetAttribute (cfg, emission_scheme, label='emission_scheme:', default='ginoux', __RC__) self%emission_scheme = ESMF_UtilStringLowerCase(trim(emission_scheme), __RC__) @@ -171,11 +167,21 @@ contains ! read FENGSHA-specific parameters ! -------------------------------- - if (self%emission_scheme == 'fengsha') then - call ESMF_ConfigGetAttribute (cfg, self%alpha, label='alpha:', __RC__) - call ESMF_ConfigGetAttribute (cfg, self%gamma, label='gamma:', __RC__) - call ESMF_ConfigGetAttribute (cfg, self%kvhmax, label='vertical_to_horizontal_flux_ratio_limit:', __RC__) - end if + select case (self%emission_scheme) + case ('fengsha') + call ESMF_ConfigGetAttribute (cfg, self%alpha, label='alpha:', __RC__) + call ESMF_ConfigGetAttribute (cfg, self%gamma, label='gamma:', __RC__) + call ESMF_ConfigGetAttribute (cfg, self%kvhmax, label='vertical_to_horizontal_flux_ratio_limit:', __RC__) + case ('k14') + call ESMF_ConfigGetAttribute (cfg, self%clayFlag, label='clayFlag:', __RC__) + call ESMF_ConfigGetAttribute (cfg, self%f_swc, label='soil_moisture_factor:', __RC__) + call ESMF_ConfigGetAttribute (cfg, self%f_scl, label='soil_clay_factor:', __RC__) + call ESMF_ConfigGetAttribute (cfg, self%uts_gamma, label='uts_gamma:', __RC__) + case ('ginoux') + ! nothing to do + case default + _ASSERT_RC(.false., "Unallowed emission scheme: "//trim(self%emission_scheme)//". Allowed: ginoux, k14, fengsha", ESMF_RC_NOT_IMPL) + end select

@rmontuoro Done and tested as zero-diff to ginoux and k14. Please re-test with fengsha

ESMF/GOCART2G_GridComp/DU2G_GridComp/DU2G_StateSpecs.rc

mathomp4 · 2021-08-06T12:29:10Z

This looks like we are close. If @rmontuoro can agree with this PR. We can pull it in as well as #71 (which doesn't affect ginoux or k14) and then make our pre-release.

rmontuoro

This revision runs with UFS but changes are recommended.

rmontuoro · 2021-08-06T14:56:58Z

Process_Library/GOCART2G_Process.F90

@@ -2392,23 +3549,28 @@ subroutine Aero_Compute_Diags (mie_table, km, klid, nbegin, nbins, rlow, rup, ch
    end if !present(extcoef)...

   if( (present(exttau) .and. associated(exttau)) .or. &


@tclune - is this valid Fortran?

if (present(a) .and. associated(a)) then

Should this be:

if (present(a)) then if (associated(a)) then

@rmontuoro I talked with @tclune and he agrees this is an issue. For now we will approve (since Intel is at least "working") and merge what is here but I'll make an issue to fix this. I don't want to try and untangle this:

GOCART/Process_Library/GOCART2G_Process.F90

Lines 3551 to 3558 in 58af774

if( (present(exttau) .and. associated(exttau)) .or. &

(present(stexttau) .and. associated(stexttau)) .or. &

(present(scatau) .and. associated(scatau)) .or. &

(present(stscatau) .and. associated(stscatau)) .or. &

(present(exttau25) .and. associated(exttau25)) .or. &

(present(exttaufm) .and. associated(exttaufm)) .or. &

(present(scatau25) .and. associated(scatau25)) .or. &

(present(scataufm) .and. associated(scataufm)) ) then

on a Friday afternoon!

Issue #72 created

Actually, this is not quite as bad as I first thought. Can be converted to 2 conditionals:

if (present(exttau) .and. present(stexttau) .and. ...) then if (associated(exttau) .and. associated(stexttau) .and. ...) then ...

rmontuoro · 2021-08-06T15:13:37Z

ESMF/GOCART2G_GridComp/GOCART2G_GridCompMod.F90


 !   Set wavelengths in universal config

    call MAPL_ConfigSetAttribute (cf, self%wavelengths_profile, label='wavelengths_for_profile_aop_in_nm:', __RC__)
    call MAPL_ConfigSetAttribute (cf, self%wavelengths_vertint, label='wavelengths_for_vertically_integrated_aop_in_nm:', __RC__)
-    call MAPL_ConfigSetAttribute (cf, wavelengths_diagmie, label='aerosol_monochromatic_optics_wavelength:', __RC__)
+    call MAPL_ConfigSetAttribute (cf, wavelengths_diagmie, label='aerosol_monochromatic_optics_wavelength_in_nm_from_LUT:', __RC__)


wavelengths_diagmie is still declared as integer:

GOCART/ESMF/GOCART2G_GridComp/GOCART2G_GridCompMod.F90

Line 113 in 58af774

integer, allocatable, dimension(:) :: wavelengths_diagmie

which is inconsistent with self%wavelengths_profile and self%wavelengths_vertint (real).

I would still recommend including changes from PR #65 into this branch to address this issue and streamline the code. For example, conversion from nm to m should be done only once (see https://github.com/GEOS-ESM/GOCART/pull/65/files#diff-682724d291af0ad213c4e1b2ffa97156de4989543290b60bf362adfa98e3d8da):

GOCART/ESMF/GOCART2G_GridComp/GOCART2G_GridCompMod.F90

Lines 156 to 162 in b9fa4c6

! Convert input wavelengths from nm to m for internal use

self%wavelengths_profile = 1.e-09 * self%wavelengths_profile

self%wavelengths_vertint = 1.e-09 * self%wavelengths_vertint

wavelengths_diagmie = 1.e-09 * wavelengths_diagmie

! Set wavelengths in universal config

@rmontuoro I guess they are declared integers because they are integers in the file:

GOCART/ESMF/GOCART2G_GridComp/GOCART2G_GridComp.rc

Line 39 in 58af774

aerosol_monochromatic_optics_wavelength_in_nm_from_LUT: 470 550 670 870

But as you said, the others aren't. For now, we might accept this as is and then you or I can make a (new) PR to fix it once we have a good "baseline" to compare to.

gmao-esherman and others added 30 commits March 9, 2021 09:07

Not checking rc for NIthermo. Temporary work around for bad rc value …

1467930

…from NIthermo.

added monochromatic_aerosol_optics call back for GAAS refactor

788fde0

added callback get_mixR, and cleaned up some .rc files

5444475

fixes diagnostic mie table bug

e1b0fc5

removed .swp file

179223d

removed prints in NI. Fixed RC bug in RPMARES.

0bd2568

removed diagnostic print statements

df755e7

deleted i,j declaration. fixed

4e32c3e

Merge branch 'develop' into feature/esherman/gocart2g_develop_v2

66848e9

Update CHANGELOG.md

c4084be

Update NI_GridCompMod.F90

14b4059

removed diagnostic print statements

chagned BOh to BOP

002fb9e

removed more prints

8bbfac6

fixed misspelled word. zero-diff change

f46dedd

fixed bug in get_mixR for CA2G_GridComp.

a4b0b02

generalized the get_mixR callback. It's now called get_mixRatioSum an…

2e7d9ce

…d is in Chem_AeroGeneric

New Phys for Dust and Carbons.

afa5af6

Fixed K14 dust emissions bug

7dfc688

added updated dust rc files to other emissions scenarios

0ce4dfe

Updated Chem_Settling, SU, and NI to latest physics

6a84f34

updated coefficient

d6e5a73

Merged AERO_ACI conents into AERO

7b4e11b

removed last AERO_ACI stuff

177d74a

added Chem_SettlingSimple

f46dd9d

removed .GOCART2G_Process.F90.swp

2b5019d

Updated wavelength parameters in GOCART2G_GridComp.rc

cd768f0

updated DU2G_GridComp/DU2G_GridComp_DU.rc parameter values

21aed0d

Removed print statements

fe40610

removed print

8d9ed2a

Update GOCART2G_GridComp.rc

62ee381

include BR

Rename sub-directory aerosol rc files to instance

c66c311

mathomp4 added 0 diff The changes in this pull request have verified to be zero-diff with the target branch. Non 0-diff The changes in this pull request are non-zero-diff labels Aug 4, 2021

mathomp4 self-assigned this Aug 4, 2021

mathomp4 requested review from tclune, weiyuan-jiang and a team as code owners August 4, 2021 16:16

mathomp4 requested a review from rmontuoro August 4, 2021 16:16

rmontuoro suggested changes Aug 4, 2021

View reviewed changes

tclune reviewed Aug 5, 2021

View reviewed changes

tclune previously approved these changes Aug 5, 2021

View reviewed changes

Add schemes

28e5589

mathomp4 dismissed tclune’s stale review via 28e5589 August 5, 2021 13:12

mathomp4 requested a review from tclune August 5, 2021 13:22

mathomp4 added 3 commits August 5, 2021 10:36

Add back print and check in SetServices

c317d89

You can't have multi-line asserts with GNU

4b18dde

Bring over FENGSHA settings

ae6ae43

rmontuoro suggested changes Aug 6, 2021

View reviewed changes

Add more scheme-specific blocks and statespecs

58af774

mathomp4 requested a review from rmontuoro August 6, 2021 13:34

rmontuoro suggested changes Aug 6, 2021

View reviewed changes

This was referenced Aug 6, 2021

Illegal Fortran in Process Library #72

Closed

Clean up nm to m conversion in GOCART2G #73

Closed

mathomp4 merged commit e392831 into develop Aug 6, 2021

mathomp4 deleted the feature/mathomp4/merge-develop-pcolarco-#1-gocart2g-updates branch August 6, 2021 15:54

	if( present(stexttau) .and. associated(stexttau) ) then
	if (ple(i,j,k) .le. tropp(i,j)) then
	stexttau(i,j,w) = stexttau(i,j,w) + tau
	elseif(ple(i,j,k) .gt. tropp(i,j) .and. ple(i,j,k-1) .lt. tropp(i,j)) then
	stexttau(i,j,w) = stexttau(i,j,w) + log(tropp(i,j)/ple(i,j,k-1))/log(ple(i,j,k)/ple(i,j,k-1))*tau
	endif
	endif

		@@ -2392,23 +3549,28 @@ subroutine Aero_Compute_Diags (mie_table, km, klid, nbegin, nbins, rlow, rup, ch
		end if !present(extcoef)...

		if( (present(exttau) .and. associated(exttau)) .or. &

	if( (present(exttau) .and. associated(exttau)) .or. &
	(present(stexttau) .and. associated(stexttau)) .or. &
	(present(scatau) .and. associated(scatau)) .or. &
	(present(stscatau) .and. associated(stscatau)) .or. &
	(present(exttau25) .and. associated(exttau25)) .or. &
	(present(exttaufm) .and. associated(exttaufm)) .or. &
	(present(scatau25) .and. associated(scatau25)) .or. &
	(present(scataufm) .and. associated(scataufm)) ) then

	! Convert input wavelengths from nm to m for internal use

	self%wavelengths_profile = 1.e-09 * self%wavelengths_profile
	self%wavelengths_vertint = 1.e-09 * self%wavelengths_vertint
	wavelengths_diagmie = 1.e-09 * wavelengths_diagmie

	! Set wavelengths in universal config

Handmerge feature/pcolarco/#1-gocart2g-updates into develop #68

Handmerge feature/pcolarco/#1-gocart2g-updates into develop #68

Conversation

mathomp4 commented Aug 4, 2021

mathomp4 commented Aug 4, 2021

rmontuoro left a 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

Choose a reason for hiding this comment

rmontuoro Aug 6, 2021 • 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

mathomp4 Aug 4, 2021 • edited Loading

Choose a reason for hiding this comment

rmontuoro Aug 4, 2021 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

tclune left a comment

Choose a reason for hiding this comment

mathomp4 commented Aug 5, 2021

rmontuoro left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

mathomp4 commented Aug 6, 2021

rmontuoro left a 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

rmontuoro Aug 6, 2021 •

edited

Loading

mathomp4 Aug 4, 2021 •

edited

Loading

rmontuoro Aug 4, 2021 •

edited

Loading