TUV-x photolysis #1471

fvitt · 2026-01-13T21:54:56Z

This provides the option to use the TUV-x library to compute photo-chemical rates in-line as an alternative to the traditional table look up method.

Closes #757

modified: .gitmodules modified: cime_config/buildlib modified: libraries/tuv-x (new commits)

jimmielin

Thanks @fvitt - I had some minor suggested changes that are hopefully straightforward.

jimmielin · 2026-01-21T20:09:03Z

src/chemistry/mozart/chemistry.F90

    call mee_ion_final()
    call rate_diags_final()
    call species_sums_final()
+    call tuvx_finalize()


Could maybe wrap this within if (tuvx_active) to be consistent with the other calls

jimmielin · 2026-01-21T20:11:40Z

src/chemistry/mozart/mo_gas_phase_chemdr.F90

+      !-----------------------------------------------------------------
+      !	... get calculated photolysis rates from TUV-x
+      !-----------------------------------------------------------------
+      call tuvx_get_photo_rates( state, pbuf, ncol, lchnk, zmid, zint, &


It looks like the only place tuvx_get_photo_rates uses state is to get state%ncol in line ~679, but ncol is already passed in here. Maybe state does not have to be passed in to call tuvx_get_photo_rates

jimmielin · 2026-01-21T20:13:47Z

src/chemistry/mozart/mo_gas_phase_chemdr.F90

+  subroutine gas_phase_chemdr(lchnk, ncol, imozart, q, &
                              phis, zm, zi, calday, &
                              tfld, pmid, pdel, pint, rpdel, rpdeldry, &
                              cldw, troplev, troplevchem, &
                              ncldwtr, ufld, vfld,  &
                              delt, ps, &
                              fsds, ts, asdir, ocnfrac, icefrac, &
                              precc, precl, snowhland, ghg_chem, latmapback, &
-                              drydepflx, wetdepflx, cflx, fire_sflx, fire_ztop, nhx_nitrogen_flx, noy_nitrogen_flx, &
-                              use_hemco, qtend, pbuf)
+                              drydepflx, wetdepflx, cflx, fire_sflx, fire_ztop, &
+                              nhx_nitrogen_flx, noy_nitrogen_flx, &
+                              use_hemco, qtend, pbuf, state)


These changes appear to be unnecessary as state was already previously passed in as the first argument, and was moved to the last one for some reason.

jimmielin · 2026-01-21T20:14:01Z

src/chemistry/mozart/mo_gas_phase_chemdr.F90

    use mo_chm_diags,      only : chm_diags, het_diags
    use perf_mod,          only : t_startf, t_stopf
    use gas_wetdep_opts,   only : gas_wetdep_method
+    use physics_types,     only : physics_state


This line is a duplicate of line 318 and does not need to be added again

The duplicate line has been removed

jimmielin · 2026-01-21T20:14:18Z

src/chemistry/mozart/chemistry.F90

+    call gas_phase_chemdr(lchnk, ncol, imozart, state%q, &
                          state%phis, state%zm, state%zi, calday, &
                          state%t, state%pmid, state%pdel, state%pint, state%rpdel, state%rpdeldry, &
                          cldw, tropLev, tropLevChem, ncldwtr, state%u, state%v, chem_dt, state%ps, &
                          fsds, cam_in%ts, cam_in%asdir, cam_in%ocnfrac, cam_in%icefrac, &
                          cam_out%precc, cam_out%precl, cam_in%snowhland, ghg_chem, state%latmapback, &
                          drydepflx, wetdepflx, cam_in%cflx, cam_in%fireflx, cam_in%fireztop, &
-                          nhx_nitrogen_flx, noy_nitrogen_flx, use_hemco, ptend%q, pbuf )
+                          nhx_nitrogen_flx, noy_nitrogen_flx, use_hemco, ptend%q, pbuf, state )


These changes appear to be unnecessary as state was already previously passed in as the first argument, and was moved to the last one for some reason.

jimmielin · 2026-01-21T20:35:36Z

src/chemistry/mozart/mo_tuvx.F90

+   integer, parameter :: PROFILE_INDEX_SO2          = 7 ! Sulfur dioxide profile index
+   integer, parameter :: PROFILE_INDEX_NO2          = 8 ! Nitrogen dioxide profile index


These appear to be non-public and unused, can I ask if these are intended to be not implemented?

jimmielin · 2026-01-21T20:35:51Z

src/chemistry/mozart/mo_tuvx.F90

+   ! Definition of the MS93 wavelength grid  TODO add description of this
+   integer,       parameter :: NUM_BINS_MS93 = 4
+   real(kind=r8), parameter :: WAVELENGTH_EDGES_MS93(NUM_BINS_MS93+1) = &
+      (/ 181.6_r8, 183.1_r8, 184.6_r8, 190.2_r8, 192.5_r8 /)


Missed TODO here about the description here

added a brief description

jimmielin · 2026-01-21T20:37:30Z

src/chemistry/mozart/mo_tuvx.F90

+      deallocate( cam_grids     )
+      deallocate( cam_profiles  )
+      deallocate( cam_radiators )
+      deallocate( wavelength )


Since wavelength => cam_grids%get_grid(...) I would suggest switching the deallocate order for safety

Suggested change

deallocate( cam_grids )

deallocate( cam_profiles )

deallocate( cam_radiators )

deallocate( wavelength )

deallocate( wavelength )

deallocate( cam_grids )

deallocate( cam_profiles )

deallocate( cam_radiators )

jimmielin · 2026-01-21T21:05:03Z

src/chemistry/mozart/mo_tuvx.F90

+
+      host_profile => profiles%get_profile( "temperature", "K" )
+      this%profiles_( PROFILE_INDEX_TEMPERATURE ) = this%core_%get_updater( host_profile, found )
+      call assert( 418735162, found )


Here and throughout the file, there are many calls to musica_assert::assert with seemingly random numbers for unique error codes. It seems to be consistent with what's done elsewhere in MUSICA, but I would request to at least have a comment somewhere that these numbers are just unique error codes and don't have any particular meaning other than being unique, and this is how the MUSICA library handles errors. Thanks!

added comments

jimmielin · 2026-01-21T21:16:14Z

src/chemistry/mozart/mo_tuvx.F90

+   !        ||                                     ||
+   !                                               ||
+   ! ----------------- i_int = pver
+   ! ================= i_imd = pver        ------------------ i_int = 2


typo here:

Suggested change

! ================= i_imd = pver ------------------ i_int = 2

! ================= i_mid = pver ------------------ i_int = 2

deleted: src/chemistry/mozart/mo_setz.F90 deleted: src/chemistry/mozart/mo_tuv_inti.F90 modified: src/chemistry/mozart/chemistry.F90 modified: src/chemistry/mozart/mo_gas_phase_chemdr.F90 modified: src/chemistry/mozart/mo_tuvx.F90

nusbaume

Thanks for bringing this in @fvitt! I had some cleanup requests in the python code, but otherwise only had a couple other comments. Hopefully none of my requests will cause any issues, but if you find that they do please let me know!

cime_config/testdefs/testmods_dirs/cam/carma_mixed_sulfate/user_nl_cam

cime_config/testdefs/testmods_dirs/cam/outfrq9s_waccm_ma_mam4/user_nl_cam

cime_config/buildlib

cime_config/buildnml

src/chemistry/mozart/chemistry.F90

src/chemistry/mozart/mo_jeuv.F90

modified: cime_config/buildlib modified: cime_config/buildnml modified: cime_config/testdefs/testmods_dirs/cam/carma_mixed_sulfate/user_nl_cam

cacraigucar

I would have expected an aux_cam regression test. Can you modify an existing test to turn this on (or is there one already)?

cacraigucar · 2026-02-03T18:45:52Z

.gitmodules

+[submodule "tuv-x"]
+path = libraries/tuv-x
+url = https://github.com/NCAR/tuv-x.git
+fxtag = ecf768a2b66


We need a tag for this instance

@K20shores, Can you make a new tag?

Created: https://github.com/NCAR/tuv-x/tree/v0.14.1

tag: v0.14.1

modified: src/chemistry/mozart/chemistry.F90 modified: src/chemistry/mozart/mo_jeuv.F90

cacraigucar · 2026-02-03T18:58:51Z

src/chemistry/mozart/mo_tuvx.F90

+         call assert_msg( 973680295, & ! random error number consistent with how MUSICA error handling
+            cam_config%validate( required_keys, optional_keys ), &
+            "Bad configuration for CAM TUV-x options." )


Just curious why this random number approach is being used instead of the __FILE__and __LINE__ which is being used elsewhere in CESM is not being used. That approach leads the user to the exact spot without needing to do a grep or table lookup.

This style of error reporting was implemented by @mattldawson. If we are not happy with this style in CAM code, I propose we use endrun to abort.

@cacraigucar Let me know if these asserts need to be replaced.

fvitt · 2026-02-03T20:48:13Z

@cacraigucar

TUV-x is turned on in this aux_cam test:

aux_cam: SMS_D_Ln9.f19_f19_mg17.FWma2000climo.derecho_intel.cam-outfrq9s_waccm_ma_mam4              # WACCM MA MAM4

See changes here:
https://github.com/ESCOMP/CAM/pull/1471/changes#diff-d1ec6e7c15fbf6c2148ef4c31be28506c102b8cf6811cb4a9dfc163edde61fd8

nusbaume

Thanks for resolving my concerns @fvitt!

However, looking over the code again I realized that there might be no need to create a build variable in the CMake python functions, as I believe everything can just use case.get_value(), so I have a couple remaining change requests related to that (along with a typo fix). Otherwise everything else looks good!

nusbaume · 2026-02-04T03:58:39Z

cime_config/testdefs/testmods_dirs/cam/carma_mixed_sulfate/user_nl_cam

 ubc_specifier = 'T->MSIS', 'Q->2.d-8vmr', 'CH4->2.d-10vmr', 'H->MSIS', 'N->MSIS', 'O->MSIS', 'O2->MSIS', 'H2->TGCM', 'NO->SNOE'
 tuvx_active = .true.
-tuvx_config_path = '$SRCROOT/src/chemistry/pp_waccm_ma_sulfur/tuvx_config.json'
+tuvx_config_path = '$COMPCOMP_ROOT_DIR_ATM/src/chemistry/pp_waccm_ma_sulfur/tuvx_config.json'


Typo here (COMPCOMP -> COMP).

nusbaume · 2026-02-04T04:27:57Z

cime_config/buildlib

-        args += "-DOS={} ".format(build.get_value("OS"))
-        args += "-DMACH={} ".format(case.get_value("MACH"))
+        args += f"-DCASEROOT={caseroot} "
+        args += f"-DCOMPILER={build.get_value('COMPILER')} "


Apologies for not catching this in the first review round, but I think you can just use the case.get_value method here and below, in which case you can remove the build = EnvBuild(case_root=caseroot) line above.

nusbaume · 2026-02-04T04:30:05Z

cime_config/buildlib

 # Builds the TUV-x library and updates the case variables used to set the
 # include paths and linked libraries

    build = EnvBuild(case_root=caseroot)


I think you can use case.get_value instead of build.get_value for all of the uses below, in which case I don't believe you need this line anymore.

fvitt added 2 commits January 7, 2026 10:28

TUV-x mods merged from tuvx_jun_cld branch

0ec3468

tuvx lib build/cmake updates

2511721

modified: .gitmodules modified: cime_config/buildlib modified: libraries/tuv-x (new commits)

cacraigucar added this to CAM Development Jan 20, 2026

cacraigucar requested review from jimmielin and nusbaume January 20, 2026 16:34

jimmielin requested changes Jan 22, 2026

View reviewed changes

changes for code review comments; remove unused modules

f193be9

deleted: src/chemistry/mozart/mo_setz.F90 deleted: src/chemistry/mozart/mo_tuv_inti.F90 modified: src/chemistry/mozart/chemistry.F90 modified: src/chemistry/mozart/mo_gas_phase_chemdr.F90 modified: src/chemistry/mozart/mo_tuvx.F90

fvitt requested a review from jimmielin January 29, 2026 15:48

jimmielin approved these changes Jan 29, 2026

View reviewed changes

nusbaume requested changes Jan 30, 2026

View reviewed changes

fvitt self-assigned this Feb 3, 2026

scripts clean up as suggested by reviewer

22003f5

modified: cime_config/buildlib modified: cime_config/buildnml modified: cime_config/testdefs/testmods_dirs/cam/carma_mixed_sulfate/user_nl_cam

cacraigucar requested changes Feb 3, 2026

View reviewed changes

minor code cleanup

9749a7d

modified: src/chemistry/mozart/chemistry.F90 modified: src/chemistry/mozart/mo_jeuv.F90

cacraigucar reviewed Feb 3, 2026

View reviewed changes

fvitt requested a review from nusbaume February 3, 2026 19:43

nusbaume requested changes Feb 4, 2026

View reviewed changes

		integer, parameter :: PROFILE_INDEX_SO2 = 7 ! Sulfur dioxide profile index
		integer, parameter :: PROFILE_INDEX_NO2 = 8 ! Nitrogen dioxide profile index

	! ================= i_imd = pver ------------------ i_int = 2
	! ================= i_mid = pver ------------------ i_int = 2

TUV-x photolysis #1471

Are you sure you want to change the base?

TUV-x photolysis #1471

Uh oh!

Conversation

fvitt commented Jan 13, 2026

Uh oh!

jimmielin left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

nusbaume left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

cacraigucar left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

fvitt commented Feb 3, 2026

Uh oh!

nusbaume left a comment

Choose a reason for hiding this comment