Semi-empirical versus Theoretical Stellar Population Models: a comparison with Star Clusters

Stellar population synthesis (SPS) models are a key tool for deriving the age, metallicity, radial velocity and reddening of star clusters from their integrated spectra. Using a sample of 129 star clusters with high-quality spectra, we analyze the uncertainties associated with selecting an empirical versus a theoretical stellar spectral library in the SPS models. We find that the fits from the different models agree on the goodness of fit metrics and inferred reddening. However, the derived age and metallicity can be affected by the choice of the stellar library, with synthetic libraries tending to give lower age and metallicity, especially for spectra with low SNR. Ages and reddening values from SSP-equivalent fits are consistent with the multi-population fits, however, SSP-equivalent metallicities are affected by the coarse coverage of the SPS grid in [Fe/H]. When comparing the spectral fitting results with the literature, we find that (1) all models underestimate age for old and metal-poor systems; (2) on average, SPS models based on synthetic stellar libraries better match the isochrone ages and metallicities from high-resolution stellar spectroscopy.