Synopsis: Some of the automated DASCH astrometric solutions are incorrect.
Details
Astrometric solutions for DASCH plates are derived in a three-step process:
- First, Astrometry.Net is used to obtain an ab initio solution.
- Next, the WCSTools imwcs program refines the solution based on a match to the Tycho-2 catalog.
- Finally, the Astromatic tool SCAMP solves for distortions across the plate area, which is very important thanks to the large angular sizes of many DASCH plates.
Unfortunately, this process sometimes fails, most often due to a divergence in the imwcs refinement step. Sometimes the SCAMP refinement succeeds even when the imwcs processing causes problems.
This problem is most likely to affect the plates that are largest on the sky. These are:
- The “meteor plates”, with plate scales > 900 arcsec/mm, comprising the
following series:
ai al bi fa ka kb ke kf kg kge kh meteor pz. - The “patrol plates”, with plate scales > 400 arcsec/mm, comprising the
following series:
ab ac aco ak am an ax ay bo ca darnor darsou dnb dnr dny dsb dsr dsy kc kd me
At the moment, we don't have a recommended scheme to identify bad solutions
automatically, although it is possible that some of the FITS headers inserted by
the imwcs program (e.g. WCSMATCH) could be helpful in doing so. They are
easy to identify by eye.
Plates with bad astrometric solutions should generally be treated as if they were never scanned. Derived lightcurve points, etc., should be ignored.
The solutions from Astrometry.Net are generally pretty good, but need refinement since DASCH plates are so large on the sky. If you have a specific plate for which you need to fix the WCS, you can probably derive a good solution yourself with some work. If you are thinking about doing this kind of work at scale, please mention your plans on the DASCH Astrophysics email list.