If I collect data using BVRI filters and do two color transforms, which filter pairs are most appropiate to report into the AID?
Should I index everything to V (do B-V, V-R, V-I)? or is there a better combination?
If I do B-V, V-R, and V-I, should I report all 3 V magnitudes or is the V magnitude from one of these three pairs better than the others and should be the value reported. Does this change for different spectral classes?
I want to say reporting the V magnitude from the B-V tranform be applicable in most cases, but I deal primarily with very red stars and suspect a different tranform may be better.
Please advise, thanks!
I'll give you my opinions on this, but I'll be interested to hear others.
You should only report one V magnitude. I usually observe in B, V, and I. For red stars I find that the transformed V magnitude from the V-I observations frequently is the better choice, but this depends on the color and brightness of the available comps, and "better" is often subjective.
I'll make a choice (between the V magnitude derived from B-V or V-I) depending on the number and quality of comps used, the resulting uncertainties, and how well the color bands I used match the derived V magnitudes for the check star to the published value. I try to pick a check star with a color close to the color of the target.
"Does this change for different spectral classes? "
The redder the star, I think the more likely I would be to pick the V value from V-I.
In principle, of course, reckoning V for some individual star using all three pairwise color-indices should give the same answer. There will naturally be some scatter as a result of having different data-sets in each group. It is traditional to determine transformations from adjacent filter-pairs; I would tend to pick the transformation that gives the smallest residuals. The thing to do is to try this on a standard field for three or four nights. Select some stars covering a range of colors that are among the standards, but do not use those in the solution. Thus you can see to what extent you can recover the standard-star values using the different color-pairs, and get some idea of systematic trends and whether one sort is preferred over another for different sorts of stars. I would suggest doing this exercise a few times per year so you can use fixed transformation coefficients over some interval (and not have to measure it every night), and see how they change as mirror coatings degrade, dust accumulates etc.
Not great to use B-V to transform very red stars. This has been gone over many times. Instead, use V-R or V-I. See the middle slides in https://github.com/edose/Spectral/blob/master/SAS%202014%20Presentation… .
V-I should statistically be better, but many I magnitudes in the charts are so egregiously bad (uncertainties of 0.4 mag not uncommon) that V-R frequently ends up being better when R magnitudes are available.