For a couple of years now I've been observing LPV Legacy stars with both B and V filters, primarily to support transformation of the V magnitudes. But Eric Dose has recently pointed out to me that B-V transforms are fairly inaccurate for red stars; V-I or V-R should be better. Yet there seem to be very few observations submitted in these longer-wavelength bands.
So is it important or useful to add R and I observations to my routine? Many of the LPV Legacy stars are already too bright for my 10-in Newtonian in V band, at least near peak brightness. R and I are likely to be even brighter. Plus, bracketing to get the correct exposure for 4 bands is going to take considerably more time than for just 2. Is it worth the trouble? Is there any AAVSO guidance on this issue?
Hi Nathan,
It is really hard to accurately transform LPVs. Transformation works well if the stars to be transformed look like blackbodies, and LPVs are far from that. In the B and V bandpasses, they look pretty normal, which is why I tend to use B&V when possible. By the time you get to Rc and Ic, the spectrum is cut up with multiple molecular bands, and usually dramatically brightening.
Transformation works best when you are interpolating and not extrapolating. At B&V, you can generally find Landolt standards that are redder and bluer than the stars you are working with, so the transforms cover the color range of interest. At Rc/Ic, LPVs are far redder than any Landolt standard, so you are extrapolating. In fact, Landolt measured very few intrinsically red stars, opting instead to use normal stars in reddened regions, to avoid many of the problems (including variability). So if transforming using (B-V), you get one value, and with (V-Ic) you get another value, which one is correct? Hard to say. My preference is to use B&V, and then to put a note on the observation indicating which filters were used for the transform. If the star is faint and you can't measure B with reasonable S/N, then you might want to use, say Ic. Again, just note the filters in your submittal. My $0.02 and not AAVSO guidance.
Arne
For the record, I have no recollection of claiming that B-V transforms are inherently "inaccurate" for red stars, and don't believe they are. (Let's hope nothing woeful like that ever slipped out after a conference beer somewhere.) Indeed one might suppose (Arne? help?) that B-V makes the best transform for LPV/Mira B magnitudes.
Perhaps it goes more like this: because (as Arne clearly points out) LPVs are far from black-body objects, B-V is fine for transforming LPV B mags, not a great choice for transforming LPV R and I mags (an extrapolation), and fine for LPV V mags if V-I is not available. My SAS 2014 presentation summarized some advantages of V-I for transforming V, R, and I of very red stars (V-I > +1.6), as supported by rather exhaustive spectral+flux simulation data (including effects of star temperature, emission lines, atmospheric absorption spectra, and various airmasses).
This agrees reasonably with Arne's guidance just above. And any remaining differences pale greatly compared to this: please, in general let's all transform LPV/Mira mags with some reasonable bands of choice before reporting. However imperfect, transforms are important to attempt.
Sorry, Eric, if I mis-interpreted a letter you posted in the maser campaign forum about V-R and V-I transforms being "better" than B-V. Based on Arne's and your guidance above, I'll stick with B-V transforms, which simplifies my observing procedure anyway (R and I measurements of the legacy LPVs so often saturate my camera!).
But my other question still remains. If the R and I bands of LPVs are so "contaminated" with molecular lines, is there any value to measuring and reporting them? For that matter, does anyone ever use the B magnitudes that I (and a few others) report?