While visually inspecting the Planet 9 search S/N map I came across a strange disk-shaped feature (NB: The source itself is not disk-shaped, this is caused by the Planet 9 search algorithm. See below.). Ignore the black circles - they just show the candidate detections triggered by this feature.
The likelihood search works by displacing individual time-chunks of data depending on the assumed velocity and distance of the object, and as these parameters are varied a given spot in a given time-chunk of data ends up being moved onto every location in a rough disk, with its size given by the maximum velocity considered. Hence, a disk-shaped feature in the total S/N map is the expected signature of a single time-chunk dominating all the others. I therefore suspected that something wrong was going on in a single one of those chunk-maps, so I plotted all of them, and this is the result.
There the rows correspond to pa4 f150, pa5 f150, pa6 f150, pa4 f220, pa5 f090 and pa6 f090 from the bottom to the top, while the columns correspond to individual 3-day chunks of data, with those that don't have exposure near this object pruned. Time increases from left to right. As you can see, there is nothing going on at this location until the very last chunk of data, all the way to the right, where a strong point source suddenly appears in all arrays that hit it (pa4 surely hits it too, but it must have been cut). This time-chunk covers the period 2019-11-7 00:00:00 to 2019-11-10 00:00:00, and sadly no s19 data after this hits this spot.
To investigate this further I find all TODs that hit this spot in that window. This turns out to be just 2 consecutive TODs: 1573251335 and 1573251992, which are mapped below:
| pa6 f150 | pa5 f150 | pa6 f090 | pa5 f090 | |
|---|---|---|---|---|
| 1573251335 2019-11-8 22:15-22:26 |
 |
 |
 |
 |
| 1573251992 2019-11-8 22:26-22:37 |
 |
 |
The event is visible as a red point in the center of the images. A normal point source is also visible about half a degree away in the southeasterly direction (down and left). As you can see, their elevation difference means pa6 first scans over the source, and then later, in the next TOD, pa5 scans over it. We can therefore map the two TODs together with no loss of time resolution. The result of this is shown below:
| pa6 f150 | pa5 f150 | pa6 f090 | pa5 f090 |
|---|---|---|---|
 |
 |
 |
 |
Here's the result of running a source finder on these. "trans" is the transient source in question, while "neigh" is the other nearby source, which is useful as a comparison for the spectral index or any gain issues. The columns prefixed by "d" indicate the 1 sigma uncertainty. The "t" and "dt" columns show the offset from 1573251335 in minutes when the source was hit, with "t" being the array mean and "dt" indicating the (tmax-tmin)/2 spread of hit times inside the array.
| Name | RA | dRA | Dec | dDec | array | t | dt | snr | flux | dflux | amp | damp |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ° | ° | ° | ° | min | min | mJy | mJy | mK | mK | |||
| trans | -86.1854 | 0.0003 | -49.4624 | 0.0006 | pa6 f150 | 8.5 | 3.2 | 20 | 518 | 26 | 5.67 | 0.28 |
| pa5 f150 | 16.6 | 3.0 | 33 | 1012 | 31 | 12.84 | 0.39 | |||||
| pa4 f150 | 16.6 | 3.0 | 16 | 732 | 46 | 8.63 | 0.55 | |||||
| pa6 f090 | 8.5 | 3.2 | 17 | 292 | 17 | 2.64 | 0.16 | |||||
| pa5 f090 | 16.6 | 3.0 | 31 | 616 | 20 | 6.06 | 0.20 | |||||
| pa4 f220 | 16.6 | 3.0 | 5 | 243 | 44 | 4.61 | 0.84 | |||||
| neigh | -85.7672 | 0.0007 | -49.7272 | 0.0012 | pa6 f150 | 9.5 | 3.2 | 10 | 324 | 33 | 5.55 | 0.40 |
| pa5 f150 | 17.6 | 3.0 | 8 | 239 | 29 | 3.03 | 0.36 | |||||
| pa4 f150 | 17.6 | 3.0 | 6 | 253 | 44 | 5.77 | 0.52 | |||||
| pa6 f090 | 9.5 | 3.2 | 17 | 320 | 18 | 2.89 | 0.17 | |||||
| pa5 f090 | 17.6 | 3.0 | 19 | 330 | 18 | 3.25 | 0.17 | |||||
| pa4 f220 | 17.6 | 3.0 | 2 | 104 | 50 | 1.97 | 0.96 |
Note how much brighter the transient object got in the 8 minutes between pa6 and pa5 saw it: It doubled in brightness at both f150 and f090. For comparison, the neighboring source is pretty constant, with 239 mJy only being a 1.9 sigma outlier. For comparison, the transient source's increase is significant at 12.2/12.3 sigma @ f150/f090. It is slightly worrying that the sources are at the edge of a somewhat bad-looking region in pa6, though.
The 1012 mJy flux is pretty bright - only 58 of the 19600 sources detected at more than 5 sigma in my f150 source catalog are brighter. Too bad the lack of any later TODs hitting this spot prevents us from seeing if it rose further in brightness afterwards. On the other hand, the non-detection at this spot in the point source catalog puts a 5 sigma upper bound on its mean flux in the s16-s19 period at 12.5 mJy. So this event represents at least a factor 80 brightening and probably considerably more.
I found a second event while checking one of the more promising planet 9 candidate detections. It looked less disk-like and more like the expected P9 signal.
However, the chunk time-series revealed it to be another transient, this time with a detection in two separate chunks.
Here are some few-TOD raw maps covering the this region.
| Time | pa4 f150 | pa5 f150 | pa6 f150 | pa4 f220 | pa5 f090 | pa6 f090 |
|---|---|---|---|---|---|---|
| 2018-10-10 23:07-23:25 |  |
 |
 |
 |
 |
 |
| 2018-10-11 23:28-24:09 |  |
 |
 |
 |
 |
 |
| 2018-10-12 23:38-24:53 |  |
 |
 |
 |
 |
 |
| 2018-10-13 23:07-23:20 |  |
 |
 |
 |
 |
 |
| 2018-10-13 02:51-03:24 |  |
 |
 |
 |
 |
 |
| 2018-10-14 23:07-24:03 |  |
 |
 |
 |
 |
 |
| 2018-10-15 02:41-02:52 |  |
 |
 |
 |
 |
 |
| 2018-10-15 23:28-24:40 |  |
 |
 |
 |
 |
 |
It's hard to see much directly from these maps, but by using forced photometry on the position of the source, ra = -58.0039±0.0003°, dec = 16.1639±0.0005°, I get these light curves. The scatter is a bit high, but all frequencies tell roughly the same story of a sudden rise followed by a gradual fall. The noisiest points correspond to the plots above where the source is at the very edge of the exposed area, and should probably be cut. The specrum is slighly blue, and quite blazar like. This might just be a blazar flaring up from below our detection threshold.
| f090 | f150 | f220 |
|---|---|---|
 |
 |
 |