[FEATURE REQUEST] Retire tropchem option

[msulprizio]

As originally reported in #732, the tropchem simulation in GEOS-Chem 13.0.0 - 13.1.2 is problematic. The issue seems to arise from the several uses of IF statements for the UCX option throughout the code and may be more of a science issue than a bug within the code. Previous timing tests have showed similar timing between GEOS-Chem standard simulations (i.e. chemistry in the troposphere + stratosphere) at the reduced 47 levels and the tropchem simulation (i.e. chemistry in the troposphere only, 47 levels). The GEOS-Chem Steering Committee has therefore recommended removing the tropchem simulation option from GEOS-Chem. This has the benefit of reducing the number of simulations maintained by the GCST and providing and should avoid confusion by reducing the number of options provided to users when setting up full-chemistry simulations.

As part of the tropchem simulation removal, we will run and distribute 1-month benchmark simulations comparing the standard (trop+strat chem) at 72 vs 47 levels. This will also include an analysis of model performance (timing, scalability). We will also run a complete set of integration tests to ensure removing the tropchem option does not break simulations that currently rely on this setup (e.g. TOMAS, RRTMG, nested grid simulations).

[msulprizio]

1-month benchmark simulations comparing the existing tropchem simulation (i.e. troposphere-only chemistry) with the standard simulation (i.e. tropospheric+stratospheric chemistry) have been completed. The model performance statistics and OH metrics are included below.

A complete set of benchmark plots and tables may be viewed at the links below. All simulations are compared to the trop+strat chemistry simulation at 72 levels as this is our standard, recommended GEOS-Chem simulation.

• Comparison of trop+strat chem (72L) vs trop+strat chem (47L)
• Comparison of trop+strat chem (72L) vs trop-only chem (72L)
• Comparison of trop+strat chem (72L) vs trop-only chem (47L)

[msulprizio]

Daniel Jacob wrote:

Your benchmark demonstrates the fidelity of 47L in the troposphere, but also makes me feel better about the need to retire trop-only chem. Specifically:

• Tropospheric masses for all species except Cl2O2 (a tropopause basket case) change by less than 1% when going from trop+strat 72L to trop+strat 47L, and my browsing through species distributions shows no large differences. This confirms that 47L is good for tropospheric applications.
• Trop-only chem shows ~50% differences for tropospheric halogens compared to trop+strat 72L, clearly because the influx from the stratosphere is messed up; the changes in OH and NOx are also non-trivial, at the 1-5% level. Time to retire it!

[msulprizio]

This is now resolved via PR #890.