AN IMPROVED METAL ABUNDANCE CALIBRATION FOR THE WASHINGTON SYSTEM

A revised metal abundance calibration for the Washington photometric system is presented which represents a significant improvement over previous calibrations in several respects. First, new observations of a number of field and open cluster giants allow a much more precise definition of the solar-abundance fiducial relation in the two-color diagrams from which the abundance-sensitive delta indices are derived. Second, observations of a large sample of globular cluster giants clearly demonstrate, and allow correction for, an unsuspected decrease in metallicity sensitivity for cooler giants. Third, a new abundance index, C - T1, and a new temperature index, M - T2, are introduced. The M - T2 color provides a much broader baseline than the T1 - T2 color and is thus much less susceptible to photometric errors in determinig abundances. Fourth, the total sample of field and cluster giants now available with Fe abundances derived from high dispersion spectroscopy is substantially larger than available previously, leading to a more extensive and accurate calibration. Various combinations of abundance and temperature indices are investigated. Metal abundance calibrations are presented for five such combinations over the range from [Fe/H] = + 0.5 to - 4, indicating that each is capable of deriving abundances to approximately 0.15 dex. The abundance indices vary by approximately 1 mag over this metallicity range. We confirm that the Washington system offers a unique combination of efficiency and accuracy for determining metallicity in late-type giants over the full range of stellar abundances, although the system loses sensitivity for the coolest metal-poor stars. Metallicities determined from the new calibration are given for some 50 open and globular clusters with previously published Washington photometry. The abundance scale established here for globular clusters is in good agreement with that of Zinn [ApJ, 293, 424 (1985)] and with that of Janes [ApJS, 39, 135 (1979)] for open clusters. Finally, comprehensive reddening and photometric error estimates are derived, as well as abundance sensitivities, and the suitability of the different indices to abundance determinations under various conditions are discussed. Under most circumstances, especially for CCD applications where substantial photometric errors may be present due to, e.g., aperture correction uncertainties, the best abundance index is the C - M color, using M - T2 as the temperature index. Thus, in many applications it is only necessary to observe in three filters, excluding T1 observations.