Dear Patrick,

Thanks a lot for the help and feedback, this is really appreciated.

For the second order term of the baseline, I believe one of the reason for it 
lies in our sideband suppression method. I'll now try to add that effect in the 
sensor_response variable and see if this impacts the baseline accordingly.


De : Patrick Eriksson <>
Envoyé : mardi 22 juin 2021 20:39:56
À : Sauvageat, Eric (IAP);
Cc :
Objet : Re: [arts-users] Tropospheric continuum retrieval in ARTS


> (1998)). The values of the poly_order=0 are now quite small (mostly
> lower than 0.1K), however, I still only get positive values. Is that
> something that could be expected from the way the function
> "retrievalAddPolyfit" works maybe ?

No, there is nothing that promotes positive values. But you now get so
small values that nothing to worry about.
You get very consistent values for poly_order 1 and 2, despite that the
tropospheric opacity varies quite a bit. That should be an indication on
that receiver is stable and the baseline+tropospheric part of the
retrievals is working well.

> Would you have any tips on how to implement the retrieval of a single
> scaling parameter for a full water vapor profile ? I believe this is
> more or less what the CONTABS_DO parameter was implementing according to
> the first Qpack documentation ?

This you do by setting the retrieval grid to have a single point (e.g.
500 hPa) and set the corresponding covariance matrix to have size 1x1. A
suitable value for it is maybe 1 (meaning a 100% uncertainty at all
altitudes, fully correlated).


arts_users.mi mailing list

Reply via email to