Is it expected that there would be Diasources reported for which there is no corresponding Source, even though the expected magnitude is above the “magLim” value for the visit detector?
I ran a TAP query on the data.lsst.cloud RSP to fetch diasources and sources which overlapped a given set of RA/Dec values, and 38 out of the 45 points found data in both tables but these (below) only turned up DiaSources without a corresponding Source within 4 arc minutes.
But maybe it landed in a place that was wiped out by a bright star or something? (black dots =sources, orange = the location of the DiaSource). Is there any way to investigate this using the existing sources and possibly their flags, before the visit images are available? And once the visit images are available, are there already tutorials (say for DP1) to guide how to investigate the presence or absence of sources in an expected location in a visit?
Hi Lynne, interesting issue! I’ve had a quick look and the diaSourceIds you have listed all have an ssObjectId = 5785499555373003312, which corresponds to 138P/Shoemaker-Levy 7 (1991 V2). I think it would make sense then for moving object DiaSources to not have an associated static Source.
As for available tutorials, if I have understood your issue correctly I would probably start with one of the cutout tutorials, e.g. 103.6. Cutout Exposures — DP2. As you have noted the DP2 visit images are not yet available but I suppose looking at the co-add images might at least reveal if there are any obvious static sources. I hope that helps!
Why would that be the case @jrob93? One would expect that for most Solar System objects that as long as they are not on a star or galaxy or blended too badly would show up in the source catalog (it’s how we found them before difference imaging became a technique Difference helps cut down on the sheer number of sources to attempt to link and false linkages but if it’s 5 sigma above the background in the difference imaging in most cases it should be in most cases be above the source detection threshold). We would have expect that those sources would have an SSObjectID. The question is how we do know why there isn’t a source where we expect. Like if pixels were masked out because of a saturated star is there way to know that ra and dec from the diasource is masked out in the PVI source catalog.
Hi Jamie,
Sorry for any confusion – when I said a corresponding Source, what I meant was no Source at the same ra/dec/time location as the DiaSource.
Most DiaSources do correspond to a Source, but there were some that didn’t (in the particular set of detections I was looking at, 38 had corresponding Sources but 7 did not ).
You’re right that these are detections of a SSObject, but that shouldn’t really matter for the Source/DiaSource detections.
Some of the 7 were close to the 5-sigma threshold (although I wonder if it might be interesting to look at this further, to understand what ‘5-sigma detection threshold’ means for the Sources and DiaSources), but some were DiaSources with SNR~20, like the one I asked about above.
So all right … waiting for visit images and difference images, and looking at the Coadds. I’ll checkout the tutorial you linked, thanks!
Ah yes you are correct Meg, I think I got a bit turned around thinking about co-add images…
And I think I understand you better now, Lynne, apologies for getting distracted by ssObjectId! I agree that it could be understood if there were differences between how 5-sigma is defined between science and difference visit images, I guess the sky background in the science images could affect things? But a high SNR DiaSource with no corresponding Source doesn’t quite make sense.
So I’m not sure what is going on here and I can’t think of a quick way to look into this without the visit pixels, but hopefully someone with better knowledge of the source detectection can chime in!
My best estimate at present is to make the kinds of plots I shared above (all of the sources across the detector and the distribution of their magnitudes) and then wait for the visit images.
If the cutout tutorial is applicable to the visit images when they’re available, I think that’s the extra bit I was looking for.
