How to calculate the age of air using SpeciesConc_CLOCK?

yc-chan commented 1 year ago

Name and Institution (Required)

Name: Yuk Chun Chan Institution: University of Washington

Confirm you have reviewed the following documentation

[x] Support guidelines
[x] User manuals
[x] Debugging GEOS-Chem and HEMCO errors
[x] Current and past Github issues

Description of your issue or question

I am interested in looking at the age of air diagnostics (the time elapsed since the air mass last contacts with the surface) in GCClassic simulations.

GEOS-Chem wiki has an outdated page about the AgeOfAir diagnostic. Since version 13.0.0, Met_AgeOfAir has been replaced by SpeciesConc_CLOCK (see the commit here). I try to search for the proper way to calculate the age of air from this clock tracer but have no luck. The only information I can find is this related discussion thread in the GCPy repository, but it does not describe the actual calculation.

How can I calculate the modeled age of air using SpeciesConc_CLOCK from my simulation outputs?

msulprizio commented 1 year ago

Hi @yc-chan. I have removed the outdated text on the wiki describing the AgeOfAir diagnostic. Thanks for pointing that out.

The CLOCK species in GEOS-Chem is a bit tricky because it's carried in State_Chm%Species concentration array like other GEOS-Chem species but it has non-standard units (e.g. seconds or timesteps vs ppbv). In main.F90, we call Set_Clock_Tracer each timestep and in doing so the number of seconds in that timestep is added to CLOCK for each grid box (and it's zeroed at the surface). In Interfaces/GCClassic/main.F90, the units of State_Chm%Species are kg species. The species concentrations are eventually converted to mol/mol dry air (using routines in GeosUtil/unitconv_mod.F90) for saving out to the GEOSChem.SpeciesConc netCDF file. To get the proper units back, you would need to convert from mol/mol dry air (same as kg/kg dry air) back to "kg" (really "seconds" for CLOCK) using this conversion as obtained from unitconv_mod.F90:

    !====================================================================
    !
    !  The conversion is as follows:
    !
    !   kg species(N)
    !   -----------  *  kg dry air
    !   kg dry air
    !
    !   = mass mixing ratio * dry air mass
    !
    !   = kg species(N)
    !
    ! Therefore, with:
    !
    !  AD(I,J,L)   = grid box dry air mass [kg]
    !
    ! the conversion is:
    !
    !  Species(I,J,L,N) [kg]
    !
    !    = Species(I,J,L,N) [kg/kg] * AD(I,J,L)
    !
    !====================================================================

The AD field can be obtained from the StateMet collection in HISTORY.rc.

NOTES:

In GEOS-Chem 14.2.0 and later versions, this has been completely updated. The CLOCK tracer is still available for use for now, but the TransportTracers simulation has removed it and uses the aoa (age of air) species instead. This species increases by a value of 1 on each emissions timestep in GeosCore/tracer_mod.F90.
The GCST is working on modifying how we do unit conversions so that we can skip certain species (like age of air) when doing unit conversions to avoid confusion for tracers with non-standard units or no molecular weights (as is the case for some of the passive tracers in the TransportTracer simulation).

yc-chan commented 1 year ago

Hi @msulprizio! Thank you a lot for the explanation! I still have one question about the unit conversion.

I am currently using version 13.3.2 for my simulations. I think the State_Chm%Species in that version still uses [kg/ kg dry air] as the unit (according to state_chm_mod.F90). In this case, should I convert SpeciesConc_CLOCK from [mol/ mol dry air] to [kg/ kg dry air] to retrieve age of air in [s]? In other words:

AoA [s] = SpeciesConc_CLOCK [mol/ mol dry air] * molar mass of clock [g/ mol] / molar mass of dry air [g/ mol dry air]

When I use molar mass of clock = 1 g/mol (according to species_database.yml) and molar mass of dry air = 28.6 g/mol, the numbers I get seem reasonable.

msulprizio commented 1 year ago

Yes @yc-chan That sounds correct. Thanks for following up. I will close out this issue, but please feel free to open a new one if you have other questions.

geoschem / geos-chem