2005PASP..117...37S

The results from work done to extend the Johnson-Cousins BVRI photometric standard sequence to faint levels of V∼21 mag in compact fields is presented. Such calibration and extension of sequences is necessary to fill a calibration gap if reliable photometry from modest aperture telescopes in space (e.g., the Hubble Space Telescope) or terrestrial telescopes with apertures exceeding 4 m is to be done. Sequences like the ones presented here, which cover a large range in brightness as well as color, will allow photometric calibration to be done efficiently and will also make such work less prone to systematic sources of error. Photometry of stars in approximately 10'x10' fields around three globular clusters (NGC 2419, Pal 4, and Pal 14) are presented from data acquired over several photometric nights. In each field, several stars are measured in B, V, R, and I passbands, with standard errors in the mean less than 0.015 mag from random errors, to levels fainter than V=21 mag. It is shown that when tying to the Landolt standards on the Johnson-Cousins system, standard errors in the mean from systematic errors are typically well below 0.01 mag in all four bands (except for B in NGC 2419 and R in Pal 4), thus justifying the claim that these fields have been correctly calibrated. The primary context for the work presented here is that parts of these fields were observed repeatedly by the Wide Field Planetary Camera 2 (WFPC2) on the HST, and thus these newly calibrated sequences can be used to retroactively calibrate WFPC2 at various times over its operating life. In the past, WFPC2 data have had typical photometric zero-point uncertainties of a few hundredths of a magnitude, largely due to a lack of suitable standard stars. The sequences presented here have standard errors at the 0.01 mag level. They agree at the 0.02 mag level with other extant calibrations of the targets presented here, except in the I band, where there are color-dependent deviations of up to 0.05 mag versus one other photometric sequence. There is no clear resolution of this difference: we present as much verification of the sequences presented here as possible. We argue that a very likely reason for such discrepancies is differences in the filter bandpass.: