I've been wanting to measure extinction on a regular basis both as part of all-sky photometric calibration, but also simply to measure it to look at long-term trends. (There are published UBV extinction data for Flagstaff, Arizona since 1955.) Among the questions I have include: 1) what is the 'optimal' airmass baseline to measure the extinction, 2) is there some minimum range in airmass, 3) what is the minimum color-range amongst the stars to give reasonable color-extinction terms?
A lot of papers from the single-channel photoelectric era suggest an aimass range of about 1.0, or at least that the high-airmass observations be done at sec z ~2.0. This may result simply from it having been difficult to do observations at higher airmass because of the difficulty of reaching the previewer eyepiece. Given that the usual Landolt equatorial fields are already at sec z ~1.25, the range of ~0.8 seems (to me) rather small. Various work by old-school people like Kris Serkowski and Andy Young suggest the high-airmass data be taken near airmass 2.35 = zenith distance 65 degrees. This apparently optimizes the airmass baseline while keeping the errors due to jitter in the photometry (seeing, scintillation, low-level crud) from being too bad.
On the star-colors side, one of the problems is that stellar statistics means the usual batches of Landolt fields (say) include stars only in the range 0.5 < B-V < 1.1. It seems one would really like to go from something near B-V = 0.0 out to 1.4 or 1.5 in order to get enough leverage for the color terms in the extinction (and cover the color range of standards one might get from a full set of standard star data). I've been poking around the sky (and literature) to find suitable red/blue pairs (just two stars within my ~20' CCD field), but these are uncommon. For instance a trawl through the Guide Star Photometric Catalogue shows only a few fields where reasonably well-measured stars are available --- and these usually only in B,V. I have been taking data around isolated blue or red Landolt stars (spectrophotometric standards) in hopes of calibrating other stars in those fields against other standards on many nights, but it would be nicer to have ones that are already measured.