Dear Ismail, dear @sander.niemeijer,
If I understand correctly, you are first using the python xarray module to ingest the COBRA SO2 data as well as the EMAC model data, you then create one single netcdf file which you then ingest using harp, so that you can apply the QA filter and bin?
I can think of a number of issues that may arise with this way of performing the analysis. I have in the past encountered differences in necdf files read in and created with xarray, which I am not sure harp is equipped to deal with. And it shouldn’t, just to be clear!
Can we assume that you are using the COBRA data available via the PAL system, from here S5P-PAL Data portal ?
In any case, I would suggest that you use the harp-compliant coda_fetch command to ingest the COBRA and the EMAC files separately, and to write them out as a single harp-compliant file. In this step, you should filter for qa for the COBRA data.
Then you will be able to use the harp_export commands to bin your data.
For e.g. in this example, Create time-series from OMI L3 daily average - Atmospheric Toolbox, a non-harp compliant netcdf file is read in with coda fetch and a harp-compliant file is created, as a new product. Then the binning is applied and L2 files created.
Do let me know if you wish any assistance in this matter.
Best wishes,
MariLiza
If that is the case