A recalibration of optical photometry:: Tycho-2, Stromgren, and Johnson systems

I use high-quality Hubble Space Telescope (HST) spectrophotometry to analyze the calibration of three popular optical photometry systems: Tycho-2 B-T and V-T, Stromgren uvby, and Johnson UBV. For Tycho-2, I revisit the analysis of an earlier paper to include the new recalibration of grating/aperture corrections, vignetting, and charge transfer inefficiency effects produced by the Space Telescope Imaging Spectrograph group and to consider the consequences of both random and systematic uncertainties. The new results reaffirm the good quality of both the Tycho-2 photometry and the HST spectrophotometry but yield a slightly different value for ZP(BT-VT) of 0.033 +/- 0.001 (random) +/- 0.005 ( systematic) mag. For the Stromgren v, b, and y filters I find that the published sensitivity curves are consistent with the available photometry and spectrophotometry, and I derive new values for the associated ZP(b-y) and ZP(m1). The same conclusion is drawn for the Johnson B and V filters and the associated ZP(B-V). The situation is different for the Stromgren u and the Johnson U filters. There I find that the published sensitivity curves yield results that are inconsistent with the available photometry and spectrophotometry, likely caused by an incorrect treatment of atmospheric effects on the short-wavelength end. I reanalyze the data to produce new average sensitivity curves for those two filters and new values for ZP(c1) and ZP(U-B). The new computation of synthetic U-B and B-V colors uses a single B sensitivity curve, which eliminates the previous unphysical existence of different definitions for each color. Finally, I find that if one uses values from the literature for which uncertainties are not given, reasonable estimates for these are 1%-2% for Stromgren b-y, m(1), and c(1) and 2%-3% for Johnson B-V and U-B. The use of the results in this paper should lead to a significant reduction of systematic errors when comparing synthetic photometry models with real colors and indices.