Affiliation
American Association of Variable Star Observers (AAVSO)
Fri, 05/01/2020 - 01:25
Greeting,
I have images from BSM Berry. The target star shows in a red circle instead of the normal green. This happens whether I use my defined comps file or use either the VSX or GCVS link to load it. See image below. This happens with all filters (BVRI), and whether the image is stacked or not. Images I got from BSM Berry of S Dor do not do this.
Dennis
While this non-identification may have many possible reasons, I'm going to throw out one that is definitely affecting my 40-year archive of CCD photometry. R Sct is a fairly high proper motion star, roughly 57mas/year. It is currently about an arcsec away from the position reported in VSX. I don't know the positional error box for VPHOT, but that might be a reason why it is not locating this variable. S Dor is not a high proper motion star, and so would not exhibit the same issue.
We will have proper motion problems for many variable stars, as often they are relatively close to us, and close stars tend to have more apparent motion. The image you take tonight shows the accurate position for the stars as of tonight. The catalog may represent the position when the variable was discovered, which could have been decades ago. Only if the coordinate epoch is known, and the proper motion is known, can you calculate its current position. This important step needs to be incorporated in all software. You can usually make the identification by eye, but computer programs need hard facts!
Arne
Arne said: "Only if the coordinate epoch is known, and the proper motion is known, can you calculate its current position. This important step needs to be incorporated in all software."
OK Arne, please tell me what photometry software packages correct for Proper Motion? VPhot does not because it uses only the VSX and GCVS catalog J2000 epoch for all star locations. I'm not aware of any commercial photometry software that does this. Are there any readily available catalogs that include both J2000 coordinates AND proper motion data that could be used to make such a position correction?
Ken
Actually, VSX contains proper motion information. It should be possible to retrieve the J2000 coordinates from VSX, along with its pmra and pmdec values, and apply the corrections for Jnow. It would require only a small amount of coding in VPHOT to do this.
I think you can set VizieR up to do this as well.
Other than that, I don't know of any package that applies proper motion. As we get farther from 2000.0, this will become more of a problem, even for some of the comp stars. My personal analysis software does apply proper motion corrections.
Arne
Arne:
Now that is REALLY useful to know! Do the proper motion ra and dec values exist for all targets or only those with large values. Even if true, we could assume that if it is missing, it could be ignored/assumed to be zero.
I will put that on the "todo" list. Of course, "a small amount of coding" is something I rarely believe but I know who I can ask to try. ;-)
Ken
I think the answer is yes? VSX PM values come from Gaia1 which probably has virtually all stars.
Ken
Actually, we must be very careful to know the epoch when applying proper motion.
For example, the ATLAS catalog, based on Gaia astrometry, reports RA and Dec based on year 2015.5, not on year 2000.0. Surprise!
This PM angle is interesting.
I recently uploaded some images of Wolf 359 (CN Leo), and with such a nearby star, there was nothing where CN Leo "ought" to be -- or was. There is also no star where one appears on the E scale chart, but I have not investigated this further. If its a dramatic variable that disappears, I'd expect it to be identified (?)
EDIT: mystery star does not appear in DSS (Aladin) and disappears when DSS option is checked for advanced chart options.
I was curious how these positions were updated in VPhot.
Hi Brad/Eric.
Yes, I remember the thread about CN Leo. It certainly moved FAST!
Just to clarify, VPhot does NOT correct for Proper Motion.
However, IF you select the real CN Leo on an image and add it to your saved sequence, it will remember that new position and that same sequence will pick that CN Leo position in future images. At least UNTIL it moves too far again! ;-) But of course, do not use the normal catalog menu again to create another saved sequence. It would look in the old position and not find it.
Effectively, you just selected an "unknown star" on the image and "called" it CN Leo.
Ken
PS: I'm not really too concerned in 2020 about stars that have moved since 2000 BUT it will get worse as we approach 2050! There are only a few that cause problems right now. As I tell all of my VPhot Choice Course students, one should ALWAYS look at their images to confirm that things are as expected BEFORE accepting their photometry. Yes, I understand Eric that it should be better than that BUT from having been burned on occasion by the "double trouble" campaign, I do not trust any software completely any more. I trust my eyes. However, we will try to make VPhot the first readily available photometry software to do this? Perhaps Pierre will get LesvePhotometry to do this before we do! ;-)
Dennis;
There are a couple of reasons for a red unidentified circle on VPhot Images.
1. I suspect that you can guess the first one based on the image above. If a target star is faint (i.e., fainter than the VPhot SNR setting - Sequence file and catalog import), it will not be confirmed as present and thus will be given a red circle at the assumed catalog location.
2. In the case of R Sct, notice that the red circle is not quite centered on the obvious target. That "likely" means that the reported catalog (VSX or GCVS) location does not match with the centroid of the visible R Sct.on this image AND that your search radius setting in VPhot is small enough that VPhot looks for R Sct at the reported VSX catalog location but does not find a sufficiently bright star in the search radius and places a red circle nearby to the current R Sct. This is my hypothesis? It is frankly NOT completely obvious that my observation is true since there is another target on your image that also shows poor centering but is identified correctly!
3. IF you removed all the red unidentified circles AND selected the obvious R Sct target manually and added it to your saved sequence, in the future this saved sequence should center the target properly. Try this out and see if my conclusion is correct? If not, there is an issue/bug?
Ken
Ken,
I tried your suggestion in #3. I cleared all unidentifies and manually added R Sct to my sequence. This works. But it is only a bandaid for the problem.
This exposed another problem I have not seen before. If it is related I do not know. I can not transform the R images. When I go to enter the color information the selection box to the right says N/A and will not let me select B-V, V-R, R-I, etc. So when I click on 'View Photometry Report' button the Transform Status is Not Transformed. There is no color index.
It works for BVI just fine.
Dennis
Dennis:
Share the BVRI image to MZK. Was the R image actually an SR image?
Ken
Ken,
I have shared the stacked images with you. The unstacked do the same thing.
I just got another set of R Sct images from BSM Berry. This R transforms as it should.
Berry is listed as having both an R and an SR filter. The problem images arrived as an R.
Dennis
Found the problem. There are two Berry Scopes. The newer one (ACP....) has a loaner but identical camera. When set up, it had only a subset of coeffs entered. Did not have all the necessary coeffs. So it did not work.
I corrected that by inputting the coeffs from the earlier camera (other Berry scope) BUT the same filters. So they should be close to valid. Try it now.
BTW, you should ask for a new sequence with all comps having BVRI mags. That is NOT the case now for many comps for RI mags! They are missing. This leads to inaccurate mags UNLESS you do not use those comps in your sequence. Did you see the negative mags? Target mags would be incorrect.
Ken
MZK said "3. IF you removed all the red unidentified circles AND selected the obvious R Sct target manually and added it to your saved sequence, in the future this saved sequence should center the target properly."
Super. Well, In 1970 Marvin Minsky said, “In three to eight years we will have a machine with the general intelligence of an average human being.” LOL. Here we are in 2020 and still, things that seem so simple for humans are still so incredibly difficult for computers. LOL.
Who still believes in it ???