Synthetic photometry of globular clusters: Uncertainties on synthetic colors

Abstract

Context. Synthetic photometry is a great tool for studying globular clusters, especially for understanding the nature of their multiple populations. Aims. Our goal is to quantify the errors on synthetic photometry that are caused by uncertainties on stellar and observational/calibration parameters. These errors can be taken into account when building synthetic color-magnitude diagrams (CMDs) that are to be compared to observed CMDs. Methods. We have computed atmosphere models and synthetic spectra for two stars, Pollux and Procyon, that have stellar parameters typical of turn-off and bottom red giant branch stars in globular clusters. We then varied the effective temperature, surface gravity, microturbulence, the carbon, nitrogen, and oxygen abundances, and [Fe/H]. We quantified the effect on synthetic photometry in the following filters: Johnson UBVRI and HST F275W, F336W, F410M, F438W, F555W, F606W, and F814W. We also estimated the effects of extinction, atmospheric correction, and of the Vega reference spectrum on the resulting photometry. In addition, we tested the ability of our models to reproduce the observed spectral energy distribution and observed photometry of the two stars. Results. We show that variations are generally stronger in blue filters, especially those below 4500 Å. Dispersions on synthetic colors due to uncertainties on stellar parameters vary between less than 0.01 and to 0.04 magnitude, depending on the choice of filters. Uncertainties on the zero points, the extinction law, or the atmospheric correction affect the resulting colors at a level of a few 0.01 magnitudes in a systematic way. The models reproduce the flux-calibrated spectral energy distribution of both stars well. Comparison between synthetic and observed UBVRI photometry shows a variable degree of (dis)agreement. The observed differences indicate that different reduction and calibration processes are performed to obtain respectively observed and synthetic photometry, and they call for publication of all the details of the reduction process to produce synthetic photometry at a 0.01 mag level, which is required to interpret observed CMDs.