Background
During A→D move, there will be cases where observations were taken when most antennas are in the new (or old) configuration, and the others are in the other configuration, leading to a skewed distribution of baselines. Images can have serious artifacts (typically seen in the form of bright fringes) due to the odd baseline lengths.
Note from John Tobin: if more than 3-5 antennas are outlying, the self-cal masking will likely fail. However, although it may be possible to "save" the image by flagging the outlying antennas, this is not ideal given that: (1) we don't know the observers' goals (they knew they were asking for move time so it's possible a nice image isn't what they're going for) (2) this may require flagging a large fraction of the data, and (3) it takes up Operations resources to spend time on this.
1 Comment
John Tobin
My comment regarding the flagging is that it could be done only for the imaging portion of the workflow, which would not result in the flags propagating to the calibration that would be used for restoration. However, I agree that it is not worth the effort to do the necessary special handling each of these move data sets would need. Especially since a user that requests move data probably knows what their getting into when the ask for such observations to be approved.