[BUG/ISSUE] AEIC VOC scale factors might not be right (?)

[flavioquadros]

Description of the problem

The scale factors applied to aviation VOC emissions were changed on v13.2.0 (#776), but I am a bit unsure if the added carbon correction was needed, could someone please check this issue again? I was under the impression that the HC variable in AEIC was in CH4-equivalent grams, as mentioned in the papers describing the inventory (https://doi.org/10.1016/j.trd.2013.07.001, https://doi.org/10.1016/j.atmosenv.2011.07.012). The speciation factors seem to be those adopted by the EPA (https://www.epa.gov/regulations-emissions-vehicles-and-engines/organic-gas-speciation-profile-aircraft), which are also implemented by the FAA's emissions model (https://aedt.faa.gov/Documents/AEDT3d_UserManual.pdf, item 5.1.1.8).

In GEOS-Chem, we have factor 114 (=1.16e-3) which does (g)->(kg) and the conversion of THC (total hydrocarbons, base CH4) to TOG (total organic gases, in overall mass). And then factors 101 ... 109 do the speciation on that as per the EPA profile. So, unless I am missing something, weren't the resulting numbers prior to the patch already in total kg of each species?

GEOS-Chem version

v13.2.x

[yantosca]

Thanks for writing @flavioquadros, sorry for the delay. The AEIC emissions data have units:

FUELBURN:units = "kg/m2/s" ;
NO2:units = "kg/m2/s" ;
HC:units = "kg/m2/s" ;
CO:units = "kg/m2/s" ;

so we have to scale that to the GEOS-Chem species as described in the HEMCO_Config.rc file.

#==============================================================================
# --- AEIC aircraft emissions ---
#
# Sulfur conversion factors are calculated as
#     so4 = 3 * a * b and so2 = 2 * a * (1-b)
#
# where
#     a = fraction by mass (6.0e-4) and
#     b = conversion efficiency (2.0e-2).
#
# The factors 2 and 3 come from sulfur mass conversions (96/32 and 64/32).
# All factors are taken from AEIC_mod.F, following Seb Estham's appraoch.
# Note that all emissions become multiplied by a factor of 1e-3 (except for
# the fuelburn emissions) in the original AEIC code. I'm not sure if this
# is because the netCDF data is in g instead of kg?
#==============================================================================
(((AEIC
101 AEICACET 3.693477e-3 - -  - xy unitless 1
102 AEICALD2 4.271822e-2 - -  - xy unitless 1
103 AEICALK4 2.137911e-1 - -  - xy unitless 1
104 AEICC2H6 5.214505e-3 - -  - xy unitless 1
105 AEICC3H8 7.808710e-4 - -  - xy unitless 1
106 AEICCH2O 1.230811e-1 - -  - xy unitless 1
107 AEICPRPE 1.780418e-1 - -  - xy unitless 1
108 AEICMACR 5.362609e-2 - -  - xy unitless 1
109 AEICRCHO 3.676944e-2 - -  - xy unitless 1
110 AEICNOCO 1.000000e-3 - -  - xy unitless 1
111 AEICSO2  1.176000e-3 - -  - xy unitless 1
112 AEICSO4  3.600000e-5 - -  - xy unitless 1
113 AEICBC   3.000000e-5 - -  - xy unitless 1
114 AEICHC   1.160000e-3 - -  - xy unitless 1

But these scale factors assumed that certain hydrocarbons were in the units kg C/m2/s, because older versions of GEOS-Chem carried certain species as kg C/m2/s. This changed in a late GC 12 or early GC 13 version (I forget which off the top of my head).

We then had to add the scale factors 40-49 to account for the the ratio (moles C / moles species) for those certain species:

0 AEIC_ACET $ROOT/AEIC/v2015-01/AEIC.47L.gen.1x1.nc  HC       2005/1-12/1/0 C xyz kg/m2/s ACET 114/101/40 20 1
0 AEIC_ALD2 -                                        -        -             - -   -       ALD2 114/102/41 20 1
0 AEIC_ALK4 -                                        -        -             - -   -       ALK4 114/103/42 20 1
0 AEIC_C2H6 -                                        -        -             - -   -       C2H6 114/104/45 20 1
0 AEIC_C3H8 -                                        -        -             - -   -       C3H8 114/105/46 20 1
0 AEIC_CH2O -                                        -        -             - -   -       CH2O 114/106    20 1
0 AEIC_PRPE -                                        -        -             - -   -       PRPE 114/107/49 20 1
0 AEIC_MACR -                                        -        -             - -   -       MACR 114/108/83 20 1
0 AEIC_RCHO -                                        -        -             - -   -       RCHO 114/109/84 20 1

Hope this helps! Am also tagging @msulprizio, who added the scale factors to the HEMCO_Config.rc files to convert kg C -> kg species.

[flavioquadros]

Thanks for the response!

The thing is I don't think these factors (101-109) assume the input is in kg C. They should be applied on total organic gases ("TOG") expressed as kg. This is, in turn, calculated from total hydrocarbons ("THC") expressed as equivalent kg of CH4, which is how turbojet emissions are measured for regulation. The hydrocarbon emissions on AEIC represent this THC in base CH4, even if at some point the species derived from it were held in GEOS-Chem as kg C, so I would think there should be no need to apply factors 40-46, 49, 83-84.

This assumption of (total) kg of organic gas is described in the EPA documents recommending this specific speciation profile (https://www.epa.gov/regulations-emissions-vehicles-and-engines/organic-gas-speciation-profile-aircraft). This is also briefly described in the technical manual of the FAA's emissions model, which adopts the same scheme (I had the wrong link in the previous message, it's items 5.1.1.8-9 of this pdf: https://aedt.faa.gov/Documents/AEDT3d_TechManual.pdf).

[yantosca]

Also tagging @sdeastham here who was the developer of the AEIC inventory.

[sdeastham]

Thanks @yantosca and @flavioquadros for digging into this! The original derivation of the UHC speciation factors comes from the AEDT guidance (https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.719.2090&rep=rep1&type=pdf), which itself is derived from an EPA document which can be found at https://www.epa.gov/sites/production/files/2016-09/420r09903_0.xls. Using acetone as a reference point, the factor 3.69...e-3 is shown as specifically the raw mass fraction of acetone in the exhaust. So I think @flavioquadros has it right and we don't need to adjust based on the number of carbon atoms. This may even be a case of the original code being incorrect, way back in AEIC_mod.F90 (or even beyond, in the now-ancient AEDT_mod.F!). I would not be surprised if we were emitting using the "true" weight fractions even though species like ACET were defined in kgC. Unburned hydrocarbons have generally been a minor contributor to aviation impacts so errors like this may not have been caught.

In short my instinct is that the mass fractions are likely usable without correction, but it would be worth re-deriving them to make absolutely sure. My apologies to @msulprizio who I fear I may have given the wrong guidance to, based on a historical view of the code rather than on the source data!

@flavioquadros - if you have an even more up to date speciation (since the GC mechanism has changed substantially since this was last really looked at), now might be a great time to implement it!

[yantosca]

Thanks @sdeastham and @flavioquadros. If one of you could make a PR with the proper speciations that would be much appreciated.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. If there are no updates within 7 days it will be closed. You can add the "never stale" tag to prevent the Stale bot from closing this issue.

[flavioquadros]

Thanks @sdeastham and @yantosca, sorry for taking so long to respond. I have looked here and indeed the speciation could use some species newly added to GEOS-Chem (like ethylene for example). I will send a PR with the fix on Monday.

[yantosca]

We can close this issue as PR #1067 has now been merged into the GEOS-Chem 13.4.0 development stream.