| | 8 | * work on the big 3pi dvo database. Searching for brown dwarfs, I discovered a problem with proper motions and some of the coordinates. It turned out that the supercosmos data got messed up during one of the relative astrometry analyses. the claimed position for the supercos points was far from the correct locations. Then, the proper motion analysis used these as real point (with large time offsets) and pushed the mean positions around a lot. I temporarily fixed this by re-calculating mean positions based only on the gpc1 and 2mass data, ignoring the supercos data for now. i will either need to re-ingest supercos (replacing detections as we go -- i have code for this) or reset the coordinates if possible. First, though, I need to understand why the supercos data got messed up so I can avoid repeating this problem. In the process of understanding this problem, and while investigating the proper motion measurements generally, I found an inconsistency in the way chisq values were calculated for the position-only (static solution) vs proper-motion & parallax solutions. I also found some errors with the flagging of objects when calculating the mean magnitudes -- the flags meant to distinguish the different levels of data quality in the mean mags were not correctly set. |
