Bug fixes for Prod/Loss diagnostics for Hg simulation -- Closes #1284, #1285

[yantosca]

This is the corresponding PR to issue #1284. It adds several minor fixes to the Hg added by @arifein, namely:

• Fixing code to archive the prod and loss diagnostics in GeosCore/mercury_mod.F90
• Added prod families PHg0, PHg2, PHg2Br, PHg2OH, PHg2Cl to Hg.kpp and species_database_hg.yml
• Setting the units for prod and loss diagnostics to molec cm-3 s-1 for all KPP-based simulations
• Rebuilding the Hg mechanism with KPP 2.5.0
• Adding the ProdLoss collection for Hg species to run/GCClassic/HISTORY.rc.templates/HISTORY.rc.Hg

Also note: I have updated default Hg restart file so that it now contains the PHg* species. You can download the updated file from: http://ftp.as.harvard.edu/gcgrid/data/ExtData/GEOSCHEM_RESTARTS/v2021-12/GEOSChem.Restart.Hg.20190101_0000z.nc4. This will let you start a simulation without having to change the time cycle flag for SPC_ from EFYO to EY in HEMCO_Config.rc.

Also tagging @viral211 @lfreese @hmhorow

[Jourdan-He]

Looks great!

[yantosca]

Thanks @Jourdan-He, I'll merge it into patch/14.0.1.

[yantosca]

GEOS-Chem Classic integration tests all passed except for the CO2 simulation:

Summary of test results:
------------------------------------------------------------------------------
Execution tests passed: 47
Execution tests failed: 1
Execution tests not yet completed: 0

The error in the CO2 simulation was tracked down to an incorrect entry in HEMCO_Config.rc for CEDS_CO2_SHP, and has now been fixed in #1356.

[jennyfisher]

Hi @yantosca @msulprizio - I'm trying to get the version of GCClassic with all of these Hg simulation fixes. I can't seem to figure out which version this is or what I need to pull. I've cloned 14.0.0-rc3 (I think? Whatever the current main version is). I looked for a branch called patch/14.0.1 or feature/HgFamilies but I couldn't find either of these.

Is there a quick and easy way to get a version with all the Hg updates? Thanks!

[yantosca]

Hi @jennyfisher, you can take the dev branch and I think that will have the fixes. We are very close to releasing 14.0.0 (once the benchmarks are approved).

[jennyfisher]

Thanks @yantosca - I have pulled the dev branch and want to double-check a few things with you, @arifein, and @viral211.

1. In [BUG/ISSUE] Hg simulation in version 13.4 missing land and ocean emissions #1285, Viral noted that the JBrO, JNO2, Br_GC and BrO_GC fields are no longer used. However these still appear in the HEMCO_Config.rc. Should we remove them from the file for clarity (and to provide people downloading unnecessary files)?

2. In the 14.0.0-rc3 master branch, all the AERO_ fields for Hg2 partitioning point to SpeciesConc files in $ROOT/GCClassic_Output/14.0.0/, whereas in the dev branch this points to files in $ROOT/GCClassic_Output/13.0.0/. Is that intentional (i.e. we switched back to the older v13 files), or an oversight, or will this be switched back when the new 14.0.0 benchmark is complete?

3. Most important: In the OXIDANT_FIELDS section of HEMCO_Config.rc, the GLOBAL_ fields all point to $ROOT/MERCURY/v2022-04/Oxidants_for_Hg_sim_2016.nc4 which only contains fields for 2016. However, we also have all of these species in the $ROOT/GCClassic_Output/ SpeciesConc files for all years from 2010-2019. Is there any reason not to update the Hg simulation to use these year-specific oxidant fields?

Thanks!
Jenny

[arifein]

Hi @jennyfisher, regarding 1.: It worked for me to run v14 with these lines deleted from HEMCO_Config.rc, so I would also be on board with this.
For 2. + 3., the others would know better. I am not up to date with how different the Br fields are in the Oxidants_for_Hg_sim_2016.nc4 file vs. the stored v14 GCClassic Output. If these oxidant fields are quite different, it might require a re-calibration of the Hg chemical scheme.

[yantosca]

@jennyfisher @arifein: We are awaiting the 14.0.0 10 year benchmark and would use those oxidant fields. We might have those available before 14.1.0 ships. I can change it to the 13.0.0 oxidant fields for now.

I can also make sure that the HEMCO_Config.rc.Hg template file is updated accordingly.

We are a bit stretched thin right now due to circumstances beyond our control. Thanks for your patience as we sort this out.

[jennyfisher]

Thanks both. In this case, sounds like we should remove the extra lines from HEMCO_Config.rc (point 1 but leave all else as is for now. Would be interested in opinion of @viral211 on point 3.

Bob - no rush, and the 13.0.0 fields are the ones in use now so I don't think you need to change that one until the 10-year benchmark is complete. Thanks!

[viral211]

@jennyfisher @arifein @yantosca: The Hg2 gas-particle partitioning no longer uses the AERO_* fields; the aerosol fields are included in the GLOBAL_* fields. These fields can also be removed from HEMCO_Config.rc
The Br fields in version 14.0 will be very different from version 13.0 since sea salt aerosol debromination has been turned off in the newer versions. @arifein is right that using these fields will need recalibration of the reduction rates.
Also, the mechanism expects the oxidant fields to be in molec/cm3, so we won't be able to directly point to the SpecicesConc fields in the GCClassic output. Maybe HEMCO can handle the unit conversion, but I am not sure about it. @yantosca would know.

[jennyfisher]

Thanks @viral211

So for now we should remove the Br fields noted above and the AERO_ fields from the HEMCO_Config.rc, and leave all other fields (e.g. oxidants) pointing to the 13.0.0 versions.

@yantosca - let us know if this makes sense or if you need more clarity!

[yantosca]

This should be all done on PR #1367 now.