Oddities in the ACTRIS-EBAS data files

[jgriesfeller]

Discussion to document oddities in the ACTRIS-EBAS data files.
From email with Paul:
Hi Paul,

while trying to match the the data found in the API and data files to
what I actually need to get to the pyaerocom variable, I found
something that might be wrong:

ncdump -h 'https://thredds.nilu.no/thredds/dodsC/ebas_doi/NV/JA/NW/NVJA-NWRD.nc'
| grep units | grep sulphate
        sulphate_corrected_ug_S_per_m3:units = "ug/m3" ;
        sulphate_corrected_ug_per_m3:units = "ug/m3" ;

I would have expected the unit for the variable
sulphate_corrected_ug_S_per_m3 to be ug S/m3.
Is that a bug or a feature?
standard_names are these:

jang@pc6001:~/tmp/ACTRIS$ ncdump -h
'https://thredds.nilu.no/thredds/dodsC/ebas_doi/NV/JA/NW/NVJA-NWRD.nc'
| grep standard_name | grep sulphate
        sulphate_corrected_ug_S_per_m3:standard_name = "mass_concentration_of_sulphate_corrected_for_seaspray_expressed_as_sulphur_in_pm10_in_air";
       sulphate_corrected_ug_per_m3:standard_name = "mass_concentration_of_sulphate_corrected_for_seaspray_in_pm10_in_air";

The background is that I thought I might need to match the unit to what I want.
For instance here:

jang@pc6001:~/tmp/ACTRIS$ ncdump -h
https://thredds.nilu.no/thredds/dodsC/ebas_doi/JX/4R/WS/JX4R-WSAT.nc |
grep standard_name | grep sulphate
        sulphate_total_ug_S_per_m3:standard_name = "mass_concentration_of_sulphate_total_in_pm2p5_in_air" ;
        sulphate_total_ug_per_m3:standard_name = "mass_concentration_of_sulphate_total_in_pm2p5_in_air" ;

both units use the same standard_name, but the units are different:

jang@pc6001:~/tmp/ACTRIS$ ncdump -h
https://thredds.nilu.no/thredds/dodsC/ebas_doi/JX/4R/WS/JX4R-WSAT.nc |
grep units | grep sulphate
        sulphate_total_ug_S_per_m3:units = "ug S/m3" ;
        sulphate_total_ug_per_m3:units = "ug/m3" ;

So in the first example matching standard_names works for my purpose,
but in the second example I have to match the unit in addition.
Stuff like this is not helping in getting an easy to maintain reader
to work. It would be good to have one defined way to find the right
variable

Jan

[jgriesfeller]

One more thing:
As you know, I have to find a way to search the API for all of our aerocom variable names. So I basically need to find a matching ACTRIS vocabulary for our pyaerocom vocabulary.
For ozone we usually use the unit ppb, but we also have ug m-3.
Looking at the ACTRIS vocabulary, I wonder, why I need to ask the API for ozone mass concentration for both and then select the right netcdf variable while the ACTRIS variable ozone amount fraction seems at first glance a better match to get the ozone data in ppb.
The thing is that the API returns some data asking for ozone amount fraction, but that refers to some French tar files and not to actual EBAS data.
Is that a bug or a feature?