I’ve been using MPO Canopus for photometry for some time, but recently needed to use VPHOT while learning how to transform my magnitudes. I’ve noticed something that bothers me. I usually use APASS mags whenever possible. MPO Canopus will usually find APASS mags for comp stars in my images. I can also find them using the Aladin program. The values from Aladin agree with those found in Canopus. I recently processed a short time series of images using both Canopus and VPHOT. To my surprise the magnitudes were quite different (~0.6 mag). I had carefully created a sequence in VPHOT with the same comps I was using in Canopus. Looking for the cause of this problem, I found the the APASS mags for the comp stars I used in Canopus were different from those listed in VPHOT. I’ve checked a couple of other fields and found similar discrepancies. These are all supposed to be APASS mags. It seems to me that something is wrong here. Can anyone help me find out what what is happening here?
Allen:
Canopus must use/query an older release of APASS (DR6 vs DR10?). I doubt (I’m sure) the majority of the comps differ by that much (0.6 mag?) but there are a few comps in the APASS database that have large random errors and such issues!
Ken
Ken,
Thanks for responding. I checked, and Canopus is using APASS DR9. VizieR also yields results from APASS DR9. I searched for APASS DR10, and found it at APASS DR10 Download | aavso. The DR10 values for the LS Cet field agree with those found in VPHOT. I didn’t bother to check the other two fields I had looked at. I also found that the uncertainties on the DR10 mags are very large, averaging 1.01 compared to 0.03 for DR9.
Canopus (APASS DR9) VPHOT (APASS DR10)
comp V unc_V B unc_B V unc_V B unc_B
GSC 632-0014 1 11.088 0.063 11.713 0.020 11.576 0.978 12.203 0.944
GSC 632-0476 2 10.755 0.016 11.692 0.017 11.372 1.042 12.362 1.001
GSC 632-1044 3 10.511 0.051 11.099 0.018 10.994 0.947 11.704 1.005
GSC 632-0203 4 12.226 0.045 13.191 0.062 12.854 1.006 13.775 0.996
GSC 632-0500 5 12.542 0.000 13.210 0.005 13.352 1.094 13.996 1.091
This raises some questions in my mind. First of all, the large uncertainties do not give me a great deal of confidence in the DR10 values. Secondly, if we are to trust the DR10 values, what does that mean about all of the photometry we have submitted to date based on the older data releases? I would expect that a new data release would be a refinement on the old values. This looks more like a revolution.
Allen
Allen:
I see what you have tabulated above and can only hypothesize that something is wrong with the query/arithmetic. The differences in magnitude between releases is several tenths mag and the errors of about 1 are not really believable.
I have used vphot apass mags for several years and releases and do not remember seeing such odd results. Could you share a few of these images in vphot to mzk, sgeo and hqa?
Ken
Ken,
Thanks for your help. The uncertainties I listed are from the APASS DR10 site.
I have shared several images, so you can look at all three fields. I was specifically trying to see if I could match VPHOT with Canopus. Consequently, I prepared special sequence files for VPHOT with the same comp stars I used in Canopus. I tried to share the appropriate sequences with you as well.
Allen
It is worth having a look at the details shown under the AAVSO Web page for APASS. Notably, for DR10, it sez: “The photometric errors appear to be worse in DR10 than in DR9.” The implication is to use DR9. Arne Henden has suggested to me that among the issues is that the reduction scheme didn’t allow them to separate out non-photometric nights very well from the genuinely clear ones. There are also problems related to zero-point gradients where fields overlap. To see some of the problems, one needs only to examine APASS values for Landolt standards, where there are often thousands of APASS observations on hundreds of nights. Sometimes there are multiple entries from the various telescopes. In some areas APASS DR9 is fine, though kinda soft with 0.03 - 0.05 mag scatter; in other areas it is poor.
For the ATLAS ‘refcat2’, John Tonry et al re-reduced the APASS Sloan g,r,i data taken from CTIO:
…and describe the 3x improvement in the results, along with going 2 magnitudes fainter than the DR9 reductions (see sections 2.2 and 3.3 of the text). One can hope that with the appearance of GAIA3, SkyMapper4, and other improvements, there’ll be a ‘refcat3’ or 4 at some point.
I tend to compare various sources of V mags (say) to look for consistency. Additional catalogues include the venerable ASAS-3 (which also has zero-point funnies), as well as published photometry for specific targets, Peter Stetson’s files, and files accessible through the CDS VizieR catalogue-query utility. Meanwhile, I take many standard fields in hopes of improving the situation.
The whole business of sky-wide multi-color photometric calibration at the couple-percent level in intermediate brightness ranges is not a solved problem.
\Brian
Folks, When using the latest APASS DR10 data for comps, suggest double checking mag;err values. I’ve had several recent examples of 2+ magnitude errors in the DR10 I and R filter mags when compared to the DR9 data.
Allen:
I got your images and sequences. I checked LS Cet images. Yes, your observations are correct. Clearly, some of the APASS DR10 mags in this field have problems! I always check them, especially errors (~1!), and discard them. It would be interesting to have the sequence team comment on how often they run into this problem and need to ignore such APASS comps. They did identify and provide a relatively consistent set of DR10 comps for the available AAVSO comps. It almost looks like there are two separate sets of APASS comps for this field in DR10?
So as others have mentioned, personally selected APASS comps must be checked for those with ‘very large/unusual’ reported errors and excluded from consideration. Errors around 0.04 are normal.
Ken
Ken,
Apparently the issue APASS mags in various data releases is not a simple one. I ran across this because I was trying to compare VPHOT with Canopus. Neither program tells the user which APASS data release it is using, nor does either program give the user a choice. The uncertainties in the comp mags are not presented. In Canopus, I can manually enter comp mags if necessary. For example, sometimes, the astrometry works better with the MPOSC3 catalog. In that case, I can start with MPOSC3 and then manually input the APASS DR9 mags from the Aladin program. I’m concerned that many VPHOT users (like me) have been taking the comp mags determined by the program to be “gospel”. It is obviously more complex than that. I suggest VPHOT should present the mags along with uncertainties and allow users to select a different APASS release if desired.
Allen
You need to beware of some of those DR10 values, some of which have unacceptably high errors. Best thing to do is submit a request to the sequence team. We often produce working sequences within hours of a submission.
Mike Poxon, sequence team
Thanks to everyone who responded to this topic. I have a related question. My CCD camera was recently serviced at SBIG, and I have taken some new data of the NGC 7790 field. I assume that the AAVSO Standard Star mags are OK. What are the errors on the standard mags?
Allen
Allen:
You should do a search of Standard Stars in aavso web site to find the full list of standard fields and their magnitudes/errors. The source for the standard fields magnitudes is not APASS but a different Henden analysis (footnote 10). The quoted errors are about 0.02-0.04, which is standard deviation not standard error of mean. Yes, the data are OK! 
Ken
Ken,
Thanks. I found that clicking the Show AAVSO Sequence under the Catalogs tab brings up a page with the mags and uncertainties.
Allen