The coadds from 2 complete nights have been examined and streaks identified. only 1 short night (data were taken for only half the night) has been nearly completely analyzed so far for sources. out of 5 bright meteor streaks, the number of false sources were:
| pt sources | ext sources |
|---|---|
| 65 | 2 |
| 5 | 0 |
| 2 | 0 |
| 0 | 0 |
| 0 | 0 |
In addition, one faint streak produced one possibly false ext. source which will be followed-up when the gal postage stamps become available again from the rtb testing.
We were greatly surprised to find so many false point sources, but in fact will now explore using the point source database to find streaks, which will work if the ratio of 30 times more false point sources than false galaxies hold up!
Although the false point sources typically have j mags of 13-14 (k mags of 12.5-13), in general the psf chi-square values are null or awful. Only 6 much fainter sources, near the limits of the search, have good chi-squares.
Position plots and color-color plots of the point sources, as well as the sources themselves, are linked in the tables.
Those who are chomping at the bit to release our "largely fantastic data" might think about how many such sources as this one are waiting for us to identify them and trash them before they reach the public:
| ra | dec | j | h | k |
|---|---|---|---|---|
| 292.89591 | 32.502972 | 12.983 | 12.846 | 12.518 |
This source, like all the sources listed above, has no optical id, making this one of the first sources to be followed up by users of our catalog, and which would then result in the first external 2MASS paper giving a list of false bright sources in our first data release.
Now before anyone gets too excited about this, note again that all the bright false sources from this particular streak would be tossed by a psf chi-squared threshold. However, much further work needs to be done to find out how many would slip below that threshold, and how many true point sources would be tossed by applying that threshold.
Although the error bars on the false source rate are very large, 72 false sources in a half night equals a rate of 0.2% for the |b|>30 deg sky, assuming 4 years of 365 nights and 100 million sources for that part of the sky. The spec for unreliability is 0.05%. Again, we should easily be able to meet that spec by applying to be determined thresholds. Clearly, we need to apply such thresholds.