COMPREHENSIVE STELLAR POPULATION-MODELS AND THE DISENTANGLEMENT OF AGE AND METALLICITY EFFECTS

The construction of detailed models for intermediate and old stellar p opulations is described. Input parameters include metallicity (-2 < [F e/H] (0.5), single-burst age (between 1.5 and 17 Gyr), and initial mas s function (IMF) exponent. Quantities output include broadband magnitu des, spectral energy distributions, surface brightness fluctuation mag nitudes, and a suite of 21 absorption feature indices. The models are checked against a wide variety of available observations. Examinations of model output yield the following conclusions. (1) If the percentag e change Delta age/Delta Z approximate to 3/2 for two populations, the y will appear almost identical in most indices. A few indices break th is degeneracy by being either more abundance sensitive (Fe4668, Fe5015 , Fe5709, and Fe5782) or more age sensitive (G4300, H beta, and presum ably higher order Balmer lines) than usual. (2) Present uncertainties in stellar evolution are of the same magnitude as the effects of IMF a nd Y in the indices studied. (3) Changes in abundance ratios (like [Mg /Fe]) are predicted to be readily apparent in the spectra of old stell ar populations. (4) The I-band flux of a stellar population is predict ed to be nearly independent of metallicity and only modestly sensitive to age. The I band is therefore recommended for standard candle work or studies of M/L in galaxies. Other conclusions stem from this work. (1) Intercomparison of models and observations of two TiO indices seem to indicate variation of the [V/Ti] ratio among galaxies, but it is n ot clear how this observation ties into the standard picture of chemic al enrichment. (2) Current estimates of [Fe/H] for the most metal-rich globulars that are based on integrated indices are probably slightly too high. (3) Colors of population models from different authors exhib it a substantial range. At solar metallicity and 13 Gyr, this range co rresponds to an age error of roughly +/-7 Gyr. Model colors from diffe rent authors applied in a differential sense have smaller uncertaintie s. (4) In the present models the dominant error for colors is probably the transformation from stellar atmospheric parameters to stellar col ors. (5) Stellar B - V is difficult to model, and current spreads amon g different authors can reach 0.2 mag. (6) If known defects in the ste llar nux library are corrected, the population model colors of this wo rk in passbands redder than U would be accurate to roughly 0.03 mag in an absolute sense. These corrections are not made in the tables of mo del output.