Moving Targets

The handling of moving targets is peculiar. When observed at different dates, the pointing and phase centers need to follow the target. This can be mistaken as a Mosaic, although the intent is to image a single field centered on the target.

UVFITS handles that through the PMRA and PMDEC columns in the AIPS_SU tables of the UVFITS files. Documentation in https://library.nrao.edu/public/memos/aips/memos/AIPSM_117.pdfAips Memo 117 () indicates that the unit is Degree par Day, but experience shows that CASA writes that in Radians per Second.

To make things worse, CASA and IMAGER do not handle the information in the same way. On one hand, CASA reports apparent positions at the epoch of observations (which thus differ when the same source is observed in different configurations¹⁵). On the other hand, IMAGER refers the position to a single epoch, normally 2000.0 (although a different reference could be used in principle), and a single set of proper motions, ensuring consistency.

UVFITS files can thus be mistaken as “tiny” mosaics. A non-zero proper motion could indicate that we are in this “false” mosaic case.

There are two ways to process such images, and both involve ignoring the apparent mosaic.

1. Ignoring the proper motion information Imaging can then be done as usual using UV_MAP provided the apparent motion results in position offsets below MOSAIC_TOLERANCE. The latter can be adjusted if needed to ensure this condition is met.

   To make sense, all data recovered from UVFITS have the same proper motion, except for observations of a solar system object where the orbits cannot be simply represented by proper motions.

   The drawback of this approach is that the epoch of the coordinates is arbitrary: it depends on the first date found in processing. Note that this is the approach used by CASA.

2. Correcting for proper motions to a known epoch. In IMAGER one can precess back to a common epoch by using the PROPER_MOTION command. The advantage is to have a clear absolute position information.

   Caution: this should be used on UV tables that are consistent in terms of positions. As of today (Sep-2024) this is unclear if the original UVFITS files has observations with different proper motion information.