Retire CO2, CH4, and tagged CO simulations

[msulprizio]

Name and Institution (Required)

Name: Melissa Sulprizio
Institution: Harvard

Describe the update

The CO2, CH4, and tagged CO simulation options have been removed from GEOS-Chem. Instead, we now use the carbon simulation in either joint (CO2-CH4-CO-OCS) or single-species mode.

All run directory files specific to those simulations have been removed for GCClassic, GCHP, and WRF. GCHP offered a separate "CO2 w/ CMS-Flux emissions" simulation option originally added for testing the GCHP adjoint, but this is option outdated and has been removed here. (Further testing of the GCHP adjoint should utilize carbon simulation with CO2 only.)

Modules co2_mod.F90, global_ch4_mod.F90, and tagged_co_mod.F90 have been deleted. Those files contained mainly hardcoded chemistry, which is now applied in KPP and carbon_gases_mod.F90 via the carbon simulation.

Expected changes

This is a zero difference change with respect to the full-chemistry benchmark simulations. It impacts run directory creation by removing options to create run directories for the CO2, CH4, and tagCO simulations.

Related Github Issue

• closes Retire CH4, CO2, tagCO simulations in favor of carbon simulation #2387
• closes Retire simulation type "CO2 w/ CMS-Flux emissions" from GCHP run directory creation GCHP#430

Validation

As originally mentioned in #2554, 1-month benchmarks for January 2019 at 4x5 resolution were run with GCClassic. Comparisons of CO2, CH4, and CO in the "offline" specialty simulations has been compared to the single-species carbon simulations.

Performance results are outlined below:

Difference plots may be viewed here:

1. gcc_carbon-CO2only_vs_CO2sim.pdf
2. gcc_carbon-CH4only_vs_CH4sim.pdf
3. gcc_carbon-COonly_vs_tagCOsim.pdf

These plots show:

1. CO2 is identical between the CO2 simulation and CO2 in the carbon simulation
2. There are small differences in CH4 in the CH4 simulation vs the CH4-only carbon simulation due to performance of the chemistry in KPP vs in global_ch4_mod.F90
3. Some differences still exist in CO when comparing the tagCO simulation with the CO-only carbon simulation, but perhaps the user community can help sort these out.

IMPORTANT: These differences should be approved by the Carbon Cycle WG Co-Chairs (@kbowman77 and @dbajones) before merging this pull request.